20230327_MLMProject_ZC
2023-03-27
SET UP
# load packages
library(tidyverse) #data wrangling## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.0 ✔ readr 2.1.4
## ✔ forcats 1.0.0 ✔ stringr 1.5.0
## ✔ ggplot2 3.4.1 ✔ tibble 3.2.0
## ✔ lubridate 1.9.2 ✔ tidyr 1.3.0
## ✔ purrr 1.0.1
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ dplyr::filter() masks stats::filter()
## ✖ dplyr::lag() masks stats::lag()
## ℹ Use the ]8;;http://conflicted.r-lib.org/conflicted package]8;; to force all conflicts to become errorslibrary(dplyr) # data wrangling
library(ggpubr) # plotting
library(readr) # read_csv
library(psych) # psych statistics - MSSD##
## Attaching package: 'psych'
##
## The following objects are masked from 'package:ggplot2':
##
## %+%, alphalibrary(lme4) # mixed models## Loading required package: Matrix
##
## Attaching package: 'Matrix'
##
## The following objects are masked from 'package:tidyr':
##
## expand, pack, unpacklibrary(effectsize) #effect size functions##
## Attaching package: 'effectsize'
##
## The following object is masked from 'package:psych':
##
## philibrary(EMAtools) #effect size functions
library(lubridate) #time variables
library(tseries) #time series tools## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zooEMA
filename <- "20230321_EMAData.csv"
EMAdata <- read_csv(filename)## Rows: 1562 Columns: 34
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (1): date
## dbl (30): id, event, day, instance, enrollmentgroup, completed, positive_sc...
## time (3): time, time_participant, time_interval
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.filename <- "20230327_betweenday.csv"
betweenday <- read_csv(filename)## Rows: 286 Columns: 16
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (1): date
## dbl (12): id, event, day, instance, enrollmentgroup, completed, positive_sc...
## time (3): time, time_participant, time_interval
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.#transform variables to factor
EMAdata$id <- as.factor(EMAdata$id)
EMAdata$event <- as.factor(EMAdata$event)
levels(EMAdata$event) <- c("Baseline", "Session 4", "Post")
EMAdata$enrollmentgroup <- as.factor(EMAdata$enrollmentgroup)
levels(EMAdata$enrollmentgroup) <- c("Waitlist Control", "Immediate Enrollment", "Group")Descriptive Statistics
#remove group 2
EMAdata <- filter(EMAdata, enrollmentgroup != "Group")Missing Data
Missing Instances
#create dataframe of missing instances (completed = 0)
missingdata <- filter(EMAdata, completed == '0')#create table of n missing: event x group
missing_instance <- matrix(0, 4,4)
colnames(missing_instance) <- c("Baseline", "Session4", "Post", "Total")
rownames(missing_instance) <- c("WLC", "IE", "Group", "Total")missing_instance[1,1] <- length(which(missingdata$event == "baseline" & missingdata$enrollmentgroup == "0"))
missing_instance[2,1] <- length(which(missingdata$event == "baseline" & missingdata$enrollmentgroup == "1"))
missing_instance[3,1] <- length(which(missingdata$event == "baseline" & missingdata$enrollmentgroup == "2"))
missing_instance[1,2] <- length(which(missingdata$event == "session4" & missingdata$enrollmentgroup == "0"))
missing_instance[2,2] <- length(which(missingdata$event == "session4" & missingdata$enrollmentgroup == "1"))
missing_instance[3,2] <- length(which(missingdata$event == "session4" & missingdata$enrollmentgroup == "2"))
missing_instance[1,3] <- length(which(missingdata$event == "post" & missingdata$enrollmentgroup == "0"))
missing_instance[2,3] <- length(which(missingdata$event == "post" & missingdata$enrollmentgroup == "1"))
missing_instance[3,3] <- length(which(missingdata$event == "post" & missingdata$enrollmentgroup == "2"))
missing_instance[1,4] <- length(which(missingdata$enrollmentgroup == "0"))
missing_instance[2,4] <- length(which(missingdata$enrollmentgroup == "1"))
missing_instance[3,4] <- length(which(missingdata$enrollmentgroup == "2"))
missing_instance[4,1] <- length(which(missingdata$event == "baseline"))
missing_instance[4,2] <- length(which(missingdata$event == "session4"))
missing_instance[4,3] <- length(which(missingdata$event == "post"))
missing_instance[4,4] <- missing_instance[4,1] + missing_instance[4,2] + missing_instance[4,3]Missing Items
#create dataframe of missing items (completed =1, N/A on at least 1 item)
missingitems <- filter(EMAdata, completed == "1" & (is.na(positive_score) | is.na(negative_score))) #n=9#impute missing items using the average of other neg/pos items
## create function
#function inputs: ID, EVENT, INSTANCE
itemavg <- function(ID, EVENT, INSTANCE){
# row from id, event, instance
r <- which(EMAdata$id == ID & EMAdata$event == EVENT & EMAdata$instance == INSTANCE)
positems <- c()
negitems <- c()
#positive items: 1, 3, 5, 9, 10, 12, 14, 16, 17, 19
positems <- append(positems, as.integer(EMAdata[r,15])) #posq1
positems <- append(positems, as.integer(EMAdata[r,17])) #posq3
positems <- append(positems, as.integer(EMAdata[r,19]))#posq5
positems <- append(positems, as.integer(EMAdata[r,23]))#posq9
positems <- append(positems, as.integer(EMAdata[r,24])) #posq10
positems <- append(positems, as.integer(EMAdata[r,26])) #posq12
positems <- append(positems, as.integer(EMAdata[r,28])) #posq14
positems <- append(positems, as.integer(EMAdata[r,30])) #posq16
positems <- append(positems, as.integer(EMAdata[r,31])) #posq17
positems <- append(positems, as.integer(EMAdata[r,33])) #posq19
posavg <- mean(positems, na.rm=TRUE)
#negative items: 2, 4, 6, 7, 8, 11, 13, 15, 18, 20
negitems <- append(negitems, as.integer(EMAdata[r,16])) #negq2
negitems <- append(negitems, as.integer(EMAdata[r,18])) #negq4
negitems <- append(negitems, as.integer(EMAdata[r,20])) #negq6
negitems <- append(negitems, as.integer(EMAdata[r,21])) #negq7
negitems <- append(negitems, as.integer(EMAdata[r,22])) #negq8
negitems <- append(negitems, as.integer(EMAdata[r,25])) #negq11
negitems <- append(negitems, as.integer(EMAdata[r,27])) #negq13
negitems <- append(negitems, as.integer(EMAdata[r,29])) #negq15
negitems <- append(negitems, as.integer(EMAdata[r,32])) #negq18
negitems <- append(negitems, as.integer(EMAdata[r,34])) #negq20
negavg <- mean(negitems, na.rm=TRUE)
#output posavg, negavg
out <- list(r, posavg, negavg)
return(out)
} #add all positive items
posscore <- function(r){
score <- as.integer(EMAdata[r, 15]) + as.integer(EMAdata[r, 17]) +
as.integer(EMAdata[r, 19]) + as.integer(EMAdata[r, 23]) +
as.integer(EMAdata[r, 24]) + as.integer(EMAdata[r, 26]) +
as.integer(EMAdata[r, 28]) + as.integer(EMAdata[r, 30]) +
as.integer(EMAdata[r, 31]) + as.integer(EMAdata[r, 33])
return(score)
}
#add all negative items
negscore <- function(r){
score <- as.integer(EMAdata[r, 16]) + as.integer(EMAdata[r, 18]) +
as.integer(EMAdata[r, 20]) + as.integer(EMAdata[r, 21]) +
as.integer(EMAdata[r, 22]) + as.integer(EMAdata[r, 25]) +
as.integer(EMAdata[r, 27]) + as.integer(EMAdata[r, 29]) +
as.integer(EMAdata[r, 32]) + as.integer(EMAdata[r, 34])
return(score)
}#12015 session 4:8
out<- itemavg("12015", "Session 4", 8)
EMAdata[23, 17] <- out[2] #posq3
EMAdata[23, 23] <- out[2] #posq9
EMAdata[23,11] <- posscore(23) #positive score
#12015 session4: 12
out<- itemavg("12015", "Session 4", 12)
EMAdata[27, 25] <- out[3] #negq11
EMAdata[27,12] <- negscore(27) #negative score
#12017 post: 3
out <- itemavg("12017", "Post", 3)
EMAdata[123, 30] <- out[2] #posq16
EMAdata[123, 11] <- posscore(123)
#12027 baseline:2
out <- itemavg("12027", "Baseline", 2)
EMAdata[542, 28] <- out[2] #posq14
EMAdata[542, 11] <- posscore(542)
#12027 session4:12
out <- itemavg("12027", "Session 4", 12)
EMAdata[567, 22] <- out[3] #neg8
EMAdata[567, 27] <- out[3] #neg13
EMAdata[567,12] <- negscore(567)
#12033 baseline:15
out<- itemavg("12033", "Baseline", 15)
EMAdata[810, 24] <- out[2] #pos10
EMAdata[810, 11] <- posscore(810)
#12033 session4:11
out<- itemavg("12033", "Session 4", 11)
EMAdata[821, 30] <- out[2] #posq16
EMAdata[821, 11] <- posscore(821)
#12039 post:2
out<- itemavg("12039", "Post", 2)
EMAdata[1097, 31] <- out[2] #posq17
EMAdata[1097,11] <- posscore(1097)
#12039 post:3
out<- itemavg("12039", "Post", 3)
EMAdata[1098, 28] <- out[2] #posq14
EMAdata[1098, 34] <- out[3] #neg20
EMAdata[1098, 11] <- posscore(1098)
EMAdata[1098, 12] <- negscore(1098)#check missing items (completed =1, N/A on at least 1 item)
missingitems <- filter(EMAdata, completed == "1" & (is.na(positive_score) | is.na(negative_score))) #n=9#save to csv
write.csv(EMAdata, "20230411_EMAData.csv")Survey Completion
Compliance Rate
#create dataframe of id x completion
compliance_rate <- matrix(0, 35, 3)
colnames(compliance_rate) <- c("Group", "Completed", "ComplianceRate")
id_names <- c("12015", "12016", "12017", "12018", "12019", "12020","12021", "12022", "12023", "12024","12025", "12026", "12027", "12028", "12029", "12030", "12031", "12032", "12033", "12034", "12035", "12036", "12037", "12038", "12039", "12040", "12041","12043", "12044", "12045", "12048","12050", "12055","12056", "12057")
rownames(compliance_rate) <- id_names#fill matrix col 1 with enrollment group
r = 1
n = 1
for(n in 1:35){
ind <- which(EMAdata$id == id_names[n])
compliance_rate[r,1] <- EMAdata$enrollmentgroup[ind[1]]
r <- r + 1
n <- n + 1
}#fill matrix col 2 with n completed by id
r = 1
n = 1
for(n in 1:35){
compliance_rate[r,2] <- length(which(EMAdata$id == id_names[n]))
r <- r + 1
n <- n + 1
}#fill matrix col 3 with compliance rate
r = 1
n = 1
for (n in 1:35){
compliance_rate[r, 3] <- compliance_rate[r, 2]/45
r <- r + 1
n <- n + 1
}#output min and max values
max <- which(compliance_rate == max(compliance_rate[,3]), arr.ind=TRUE)
min <- which(compliance_rate == min(compliance_rate[,3]), arr.ind=TRUE)
print(paste("Maximum compliance rate:", compliance_rate[max[1]]))## [1] "Maximum compliance rate: 1"print(paste("Minimum compliance rate:", compliance_rate[min]))## [1] "Minimum compliance rate: 0" "Minimum compliance rate: 0"
## [3] "Minimum compliance rate: 0" "Minimum compliance rate: 0"
## [5] "Minimum compliance rate: 0" "Minimum compliance rate: 0"Mean (SD) Table
summarise_at(EMAdata, vars(positive_score, negative_score), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 1 × 4
## positive_score_mean negative_score_mean positive_score_sd negative_score_sd
## <dbl> <dbl> <dbl> <dbl>
## 1 22.5 15.9 7.82 4.61#group by event
EMAdata%>%
group_by(event) %>%
summarise_at(vars(positive_score, negative_score), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 3 × 5
## event positive_score_mean negative_score_mean positive_score_sd negative…¹
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 Baseline 21.5 16.0 7.10 4.73
## 2 Session 4 21.8 16.0 7.56 4.89
## 3 Post 24.4 15.7 8.43 4.15
## # … with abbreviated variable name ¹negative_score_sd#group by enrollment
EMAdata%>%
group_by(enrollmentgroup) %>%
summarise_at(vars(positive_score, negative_score), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 2 × 5
## enrollmentgroup positive_score_mean negative_score_mean positiv…¹ negat…²
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 Waitlist Control 22.8 16.0 7.97 4.23
## 2 Immediate Enrollment 22.3 15.8 7.67 4.94
## # … with abbreviated variable names ¹positive_score_sd, ²negative_score_sdNormality & Distribution
#create grouped boxplot - positive
ggplot(EMAdata, aes(x=event, y=positive_score, fill=enrollmentgroup)) + scale_fill_manual(values = c("lightblue3", "dodgerblue4")) + geom_boxplot() + xlab("event") + ylab("score") + ggtitle("PANAS Positive Score x Group")## Warning: Removed 185 rows containing non-finite values (`stat_boxplot()`).
#create grouped boxplot - negative
ggplot(EMAdata, aes(x=event, y=negative_score, fill=enrollmentgroup)) +
scale_fill_manual(values = c("lightblue3", "dodgerblue4")) +
geom_boxplot() + xlab("event") + ylab("score") +
ggtitle("PANAS Negative Score x Group")## Warning: Removed 185 rows containing non-finite values (`stat_boxplot()`).
## Affect Dynamic Variables
withinday <- EMAdata[,c('id', 'event', 'day', 'instance', 'enrollmentgroup', 'time_interval', 'positive_score', 'negative_score', "positive_SD","negative_SD")]#convert time_interval from h:m:s to minutes
withinday$time_interval_min <- hour(withinday$time_interval)*60 + minute(withinday$time_interval)
withinday <- withinday %>%
relocate("time_interval_min", .after="time_interval")# scatter plot of ASD x time interval (hour)
ggplot(withinday, aes(x = time_interval, y=positive_SD)) + geom_point(size = 0.5) +
ggtitle("Successive Difference - Positive")## Warning: Removed 406 rows containing missing values (`geom_point()`).
ggplot(withinday, aes(x = time_interval, y=negative_SD)) + geom_point(size = 0.5) +
ggtitle("Successive Difference - Negative")## Warning: Removed 406 rows containing missing values (`geom_point()`).
#ASD = positive_SD[i] / time_interval[i]
pos_ASD <- vector()
for(i in 1:1427){
pos_ASD <- append(pos_ASD, c(withinday$positive_SD[i] / withinday$time_interval_min[i]))
i = i + 1
}
head(pos_ASD)## [1] NA 0.007246377 0.000000000 0.003745318 0.024752475 NA#create positive ASD variable
withinday$positive_ASD <- pos_ASDneg_ASD <- vector()
for(i in 1:1427){
neg_ASD <- append(neg_ASD, c(withinday$negative_SD[i] / withinday$time_interval_min[i]))
i = i + 1
}
head(neg_ASD)## [1] NA 0.01449275 0.00000000 0.00000000 0.00000000 NA#create negative ASD variable
withinday$negative_ASD <- neg_ASD# scatter plot of ASD x time interval (hour)
ggplot(withinday, aes(x = time_interval, y=positive_ASD)) + geom_point(size = 0.5) +
ggtitle("Adj. Successive Difference - Positive")## Warning: Removed 413 rows containing missing values (`geom_point()`).
ggplot(withinday, aes(x = time_interval, y=negative_ASD)) + geom_point(size = 0.5) +
ggtitle("Adj. Successive Difference - Negative")## Warning: Removed 413 rows containing missing values (`geom_point()`).
betweenday <- filter(betweenday, betweenday$id != "12046")
#split betweenday df by event
betweenday_baseline <- filter(betweenday, betweenday$event == "1")
betweenday_session4 <- filter(betweenday, betweenday$event == "2")
betweenday_post <- filter(betweenday, betweenday$event == "3")#remove group 2 & 12046 (no between data)
id_names <- c("12015", "12016", "12017", "12018", "12019", "12020","12021", "12022", "12023", "12024","12025", "12026", "12027", "12028", "12029", "12030", "12031", "12032", "12033", "12034", "12035", "12036", "12037", "12038", "12039", "12040", "12041","12043", "12044", "12045", "12050", "12057")
length(id_names)## [1] 32id_group <- c("1", "1", "0", "1", "0", "1", "0", "0", "1", "1", "0", "1", "0", "0", "0", "0", "1", "1", "1", "0", "1", "0", "0", "1", "1", "0", "0", "1", "0", "1","1", "1")
length(id_group)## [1] 32Mean
# initialize MEAN matrix
ema_mean <- matrix(0, 32, 7)
colnames(ema_mean) <- c("group", "posbaseline", "possession4", "pospost", "negbaseline", "negsession4", "negpost")
rownames(ema_mean) <- id_names
for(i in 1:32){
ema_mean[i,1] <- id_group[i]
}#fill positive mean
## baseline
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_baseline$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_baseline$positive_mean[index[x]]
vec <- c(vec, n)
}
ema_mean[i,2] <- mean(vec)
}
## session4
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_session4$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_session4$positive_mean[index[x]]
vec <- c(vec, n)
}
ema_mean[i,3] <- mean(vec)
}
## post
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_post$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_post$positive_mean[index[x]]
vec <- c(vec, n)
}
ema_mean[i,4] <- mean(vec)
}#fill negative mean
## baseline
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_baseline$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_baseline$negative_mean[index[x]]
vec <- c(vec, n)
}
ema_mean[i,5] <- mean(vec)
}
## session4
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_session4$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_session4$negative_mean[index[x]]
vec <- c(vec, n)
}
ema_mean[i,6] <- mean(vec)
}
## post
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_post$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_post$negative_mean[index[x]]
vec <- c(vec, n)
}
ema_mean[i,7] <- mean(vec)
}ema_mean <- as.data.frame(ema_mean)
ema_mean$posbaseline <- as.numeric(ema_mean$posbaseline)
ema_mean$possession4 <- as.numeric(ema_mean$possession4)
ema_mean$pospost <- as.numeric(ema_mean$pospost)
ema_mean$negbaseline <- as.numeric(ema_mean$negbaseline)
ema_mean$negsession4 <- as.numeric(ema_mean$negsession4)
ema_mean$negpost <- as.numeric(ema_mean$negpost)#group by group
ema_mean%>%
group_by(group) %>%
summarise_at(vars(posbaseline, possession4, pospost, negbaseline, negsession4, negpost, ), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 2 × 13
## group posbas…¹ posse…² pospo…³ negba…⁴ negse…⁵ negpo…⁶ posba…⁷ posse…⁸ pospo…⁹
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0 21.7 23.3 23.2 16.5 15.9 16.6 5.61 6.10 6.16
## 2 1 22.3 20.7 23.9 17.3 16.0 16.2 5.77 5.30 5.87
## # … with 3 more variables: negbaseline_sd <dbl>, negsession4_sd <dbl>,
## # negpost_sd <dbl>, and abbreviated variable names ¹posbaseline_mean,
## # ²possession4_mean, ³pospost_mean, ⁴negbaseline_mean, ⁵negsession4_mean,
## # ⁶negpost_mean, ⁷posbaseline_sd, ⁸possession4_sd, ⁹pospost_sdPlots
#create id variable
ema_mean$id <- id_names
#reformat data: baseline
ema_mean_baseline <- ema_mean %>%
dplyr::select(id, group, pos = posbaseline, neg = negbaseline)
#create time variable
ema_mean_baseline$time <- as.integer(1)
ema_mean_baseline <- relocate(ema_mean_baseline, time, .before=group)#reformat data: session4
ema_mean_session4 <- ema_mean %>%
dplyr::select(id, group, pos = possession4, neg = negsession4)
#create time variable
ema_mean_session4$time <- as.integer(2)
ema_mean_session4 <- relocate(ema_mean_session4, time, .before=group)#reformat data: post
ema_mean_post <- ema_mean %>%
dplyr::select(id, group, pos = pospost, neg = negpost)
#create time variable
ema_mean_post$time <- as.integer(3)
ema_mean_post<- relocate(ema_mean_post, time, .before=group)#join predata and postdata
ema_mean2 <- merge(ema_mean_baseline, ema_mean_session4, all=TRUE)
ema_mean2 <- merge(ema_mean2, ema_mean_post, all=TRUE)
ema_mean2$id <- factor(ema_mean2$id)
ema_mean2$group <- factor(ema_mean2$group)
levels(ema_mean2$group) <- c("Control", "Treatment")
ema_mean2$time <- factor(ema_mean2$time)
levels(ema_mean2$time) <- c("baseline", "session4", "post")#plot positive means
meanpos_ggplot<- ggplot(data= ema_mean2,
aes(y=pos, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("Mean Positive Affect Across Timepoints") +
labs(y = "positive affect score", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
meanpos_ggplot## Warning: Removed 1 rows containing missing values (`geom_point()`).## Warning: Removed 1 row containing missing values (`geom_line()`).
#plot negative means
meanneg_ggplot<- ggplot(data= ema_mean2,
aes(y=neg, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("Mean Negative Affect Across Timepoints") +
labs(y = "negative affect score", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
meanneg_ggplot## Warning: Removed 1 rows containing missing values (`geom_point()`).## Warning: Removed 1 row containing missing values (`geom_line()`).
Variability (SD)
# initialize SD matrix
ema_sd <- matrix(0, 32, 7)
colnames(ema_sd) <- c("group", "posbaseline", "possession4", "pospost", "negbaseline", "negsession4", "negpost")
rownames(ema_sd) <- id_names
for(i in 1:32){
ema_sd[i,1] <- id_group[i]
}#fill positive mean
## baseline
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_baseline$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_baseline$positive_mean[index[x]]
vec <- c(vec, n)
}
ema_sd[i,2] <- sd(vec)
}
## session4
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_session4$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_session4$positive_mean[index[x]]
vec <- c(vec, n)
}
ema_sd[i,3] <- sd(vec)
}
## post
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_post$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_post$positive_mean[index[x]]
vec <- c(vec, n)
}
ema_sd[i,4] <- sd(vec)
}#fill negative mean
## baseline
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_baseline$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_baseline$negative_mean[index[x]]
vec <- c(vec, n)
}
ema_sd[i,5] <- sd(vec)
}
## session4
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_session4$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_session4$negative_mean[index[x]]
vec <- c(vec, n)
}
ema_sd[i,6] <- sd(vec)
}
## post
for(i in 1:32){
name <- id_names[i]
index <- which(betweenday_post$id == name)
vec <- c()
len <- length(index)
for(x in 1:len){
n <- betweenday_post$negative_mean[index[x]]
vec <- c(vec, n)
}
ema_sd[i,7] <- sd(vec)
}ema_sd <- as.data.frame(ema_sd)
ema_sd$posbaseline <- as.numeric(ema_sd$posbaseline)
ema_sd$possession4 <- as.numeric(ema_sd$possession4)
ema_sd$pospost <- as.numeric(ema_sd$pospost)
ema_sd$negbaseline <- as.numeric(ema_sd$negbaseline)
ema_sd$negsession4 <- as.numeric(ema_sd$negsession4)
ema_sd$negpost <- as.numeric(ema_sd$negpost)#group by group
ema_sd%>%
group_by(group) %>%
summarise_at(vars(posbaseline, possession4, pospost, negbaseline, negsession4, negpost, ), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 2 × 13
## group posbas…¹ posse…² pospo…³ negba…⁴ negse…⁵ negpo…⁶ posba…⁷ posse…⁸ pospo…⁹
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0 3.25 3.79 2.87 2.40 2.00 2.45 1.87 2.31 2.10
## 2 1 3.47 3.94 3.01 2.59 1.95 1.84 2.42 2.60 1.44
## # … with 3 more variables: negbaseline_sd <dbl>, negsession4_sd <dbl>,
## # negpost_sd <dbl>, and abbreviated variable names ¹posbaseline_mean,
## # ²possession4_mean, ³pospost_mean, ⁴negbaseline_mean, ⁵negsession4_mean,
## # ⁶negpost_mean, ⁷posbaseline_sd, ⁸possession4_sd, ⁹pospost_sdPlots
#create id variable
ema_sd$id <- id_names
#reformat data: baseline
ema_sd_baseline <- ema_sd %>%
dplyr::select(id, group, pos = posbaseline, neg = negbaseline)
#create time variable
ema_sd_baseline$time <- as.integer(1)
ema_sd_baseline <- relocate(ema_sd_baseline, time, .before=group)#reformat data: session4
ema_sd_session4 <- ema_mean %>%
dplyr::select(id, group, pos = possession4, neg = negsession4)
#create time variable
ema_sd_session4$time <- as.integer(2)
ema_sd_session4 <- relocate(ema_sd_session4, time, .before=group)#reformat data: post
ema_sd_post <- ema_mean %>%
dplyr::select(id, group, pos = pospost, neg = negpost)
#create time variable
ema_sd_post$time <- as.integer(3)
ema_sd_post<- relocate(ema_sd_post, time, .before=group)#join predata and postdata
ema_sd2 <- merge(ema_sd_baseline, ema_sd_session4, all=TRUE)
ema_sd2 <- merge(ema_sd2 , ema_sd_post, all=TRUE)
ema_sd2 $id <- factor(ema_sd2 $id)
ema_sd2 $group <- factor(ema_sd2 $group)
levels(ema_sd2 $group) <- c("Control", "Treatment")
ema_sd2 $time <- factor(ema_sd2 $time)
levels(ema_sd2 $time) <- c("baseline", "session4", "post")#plot positive means
sdpos_ggplot<- ggplot(data= ema_sd2,
aes(y=pos, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("SD of Positive Affect Across Timepoints") +
labs(y = "SD", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
sdpos_ggplot## Warning: Removed 1 rows containing missing values (`geom_point()`).## Warning: Removed 1 row containing missing values (`geom_line()`).
#plot negative means
sdneg_ggplot<- ggplot(data= ema_sd2,
aes(y=neg, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("SD of Negative Affect Across Timepoints") +
labs(y = "SD", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
sdneg_ggplot## Warning: Removed 1 rows containing missing values (`geom_point()`).## Warning: Removed 1 row containing missing values (`geom_line()`).
### Intertia (Auto-Correlation)
# create AC df
EMAdata_ac <- EMAdata[,1:14]
EMAdata_ac$positive_ac <- c(0)
EMAdata_ac$negative_ac <- c(0)
EMAdata_ac <- filter(EMAdata_ac, !is.na(positive_score))#compute AC for positive score
for(n in 1:32){ #cycle through participant ids
name <- id_names[n]
for(d in 1:9){ #cycle through days
index <- which(EMAdata_ac$id == name & EMAdata_ac$day == d)
vec <- c()
len <- length(index)
if(len !=0){#issue if there is no entry for a day number
for(i in 1:len){ #number of instances
n <- EMAdata_ac$positive_score[index[i]]
vec <- c(vec,n) #create vector of pos scores for each instance
}
acf <- unlist(acf(vec, pl=FALSE))
for(i in 1:len){ #number of instances
EMAdata_ac$positive_ac[index[i]] <- acf[i]
}
}
}
}#compute AC for negative score
for(n in 1:32){ #cycle through participant ids
name <- id_names[n]
for(d in 1:9){ #cycle through days
index <- which(EMAdata_ac$id == name & EMAdata_ac$day == d)
vec <- c()
len <- length(index)
if(len !=0){#issue if there is no entry for a day number
for(i in 1:len){ #number of instances
n <- EMAdata_ac$negative_score[index[i]]
vec <- c(vec,n) #create vector of neg scores for each instance
}
acf <- unlist(acf(vec, pl=FALSE))
for(i in 1:len){ #number of instances
EMAdata_ac$negative_ac[index[i]] <- acf[i]
}
}
}
}# initialize AC matrix
ema_ac <- matrix(0, 32, 7)
colnames(ema_ac) <- c("group", "posbaseline", "possession4", "pospost", "negbaseline", "negsession4", "negpost")
rownames(ema_ac) <- id_names
for(i in 1:32){
ema_ac[i,1] <- id_group[i]
}EMAdata_ac$positive_ac <- as.numeric(EMAdata_ac$positive_ac)
EMAdata_ac$negative_ac <- as.numeric(EMAdata_ac$negative_ac)#fill positive AC
## baseline
for(i in 1:32){
name <- id_names[i]
index <- which(EMAdata_ac$id == name & EMAdata_ac$event == "Baseline")
vec <- c()
len <- length(index)
for(x in 1:len){
n <- EMAdata_ac$positive_ac[index[x]]
if(n != "1"){
vec <- c(vec, n)}
}
ema_ac[i,2] <- mean(vec, na.rm=TRUE)
}## session4
for(i in 1:32){
name <- id_names[i]
index <- which(EMAdata_ac$id == name & EMAdata_ac$event == "Session 4")
vec <- c()
len <- length(index)
for(x in 1:len){
n <- EMAdata_ac$positive_ac[index[x]]
if(n != "1"){
vec <- c(vec, n)
}
}
ema_ac[i,3] <- mean(vec)
}## post
for(i in 1:32){
name <- id_names[i]
index <- which(EMAdata_ac$id == name & EMAdata_ac$event == "Post")
vec <- c()
len <- length(index)
if(len !=0){
for(x in 1:len){
n <- EMAdata_ac$positive_ac[index[x]]
if(n != "1"){
vec <- c(vec, n)
}
}
}
ema_ac[i,4] <- mean(vec, na.rm=TRUE)
}## Warning in mean.default(vec, na.rm = TRUE): argument is not numeric or logical:
## returning NA#fill negative AC
## baseline
for(i in 1:32){
name <- id_names[i]
index <- which(EMAdata_ac$id == name & EMAdata_ac$event == "Baseline")
vec <- c()
len <- length(index)
for(x in 1:len){
n <- EMAdata_ac$negative_ac[index[x]]
if(n != "1"){
vec <- c(vec, n)}
}
ema_ac[i,5] <- mean(vec, na.rm=TRUE)
}## session4
for(i in 1:32){
name <- id_names[i]
index <- which(EMAdata_ac$id == name & EMAdata_ac$event == "Session 4")
vec <- c()
len <- length(index)
for(x in 1:len){
n <- EMAdata_ac$negative_ac[index[x]]
if(n != "1"){
vec <- c(vec, n)
}
}
ema_ac[i,6] <- mean(vec)
}## post
for(i in 1:32){
name <- id_names[i]
index <- which(EMAdata_ac$id == name & EMAdata_ac$event == "Post")
vec <- c()
len <- length(index)
if(len !=0){
for(x in 1:len){
n <- EMAdata_ac$negative_ac[index[x]]
if(n != "1"){
vec <- c(vec, n)
}
}
}
ema_ac[i,7] <- mean(vec, na.rm=TRUE)
}## Warning in mean.default(vec, na.rm = TRUE): argument is not numeric or logical:
## returning NAema_ac <- as.data.frame(ema_ac)
ema_ac$posbaseline <- as.numeric(ema_ac$posbaseline)
ema_ac$possession4 <- as.numeric(ema_ac$possession4)
ema_ac$pospost <- as.numeric(ema_ac$pospost)
ema_ac$negbaseline <- as.numeric(ema_ac$negbaseline)
ema_ac$negsession4 <- as.numeric(ema_ac$negsession4)
ema_ac$negpost <- as.numeric(ema_ac$negpost)#group by group
ema_ac%>%
group_by(group) %>%
summarise_at(vars(posbaseline, possession4, pospost, negbaseline, negsession4, negpost, ), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 2 × 13
## group posbas…¹ posse…² pospo…³ negba…⁴ negse…⁵ negpo…⁶ posba…⁷ posse…⁸ pospo…⁹
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0 -0.152 -0.140 -0.143 -0.152 -0.140 -0.143 0.0345 0.0218 0.0258
## 2 1 -0.172 -0.149 -0.189 -0.174 -0.149 -0.189 0.0882 0.0560 0.124
## # … with 3 more variables: negbaseline_sd <dbl>, negsession4_sd <dbl>,
## # negpost_sd <dbl>, and abbreviated variable names ¹posbaseline_mean,
## # ²possession4_mean, ³pospost_mean, ⁴negbaseline_mean, ⁵negsession4_mean,
## # ⁶negpost_mean, ⁷posbaseline_sd, ⁸possession4_sd, ⁹pospost_sdPlots
#create id variable
ema_ac$id <- id_names
#reformat data: baseline
ema_ac_baseline <- ema_ac %>%
dplyr::select(id, group, pos = posbaseline, neg = negbaseline)
#create time variable
ema_ac_baseline$time <- as.integer(1)
ema_ac_baseline <- relocate(ema_ac_baseline, time, .before=group)#reformat data: session4
ema_ac_session4 <- ema_ac %>%
dplyr::select(id, group, pos = possession4, neg = negsession4)
#create time variable
ema_ac_session4 $time <- as.integer(2)
ema_ac_session4 <- relocate(ema_ac_session4 , time, .before=group)#reformat data: post
ema_ac_post <- ema_ac %>%
dplyr::select(id, group, pos = pospost, neg = negpost)
#create time variable
ema_ac_post$time <- as.integer(3)
ema_ac_post<- relocate(ema_ac_post, time, .before=group)#join predata and postdata
ema_ac2 <- merge(ema_ac_baseline, ema_ac_session4, all=TRUE)
ema_ac2 <- merge(ema_ac2, ema_ac_post, all=TRUE)
ema_ac2 $id <- factor(ema_ac2$id)
ema_ac2 $group <- factor(ema_ac2$group)
levels(ema_ac2 $group) <- c("Control", "Treatment")
ema_ac2 $time <- factor(ema_ac2$time)
levels(ema_ac2$time) <- c("baseline", "session4", "post")#plot positive means
acpos_ggplot<- ggplot(data= ema_ac2,
aes(y=pos, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("AC of Positive Affect Scores Across Timepoints") +
labs(y = "autocorrelation", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
acpos_ggplot## Warning: Removed 1 rows containing missing values (`geom_point()`).## Warning: Removed 1 row containing missing values (`geom_line()`).
#plot negative means
acneg_ggplot<- ggplot(data= ema_ac2,
aes(y=neg, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("AC of Negative Affect Scores Across Timepoints") +
labs(y = "autocorrelation", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
acneg_ggplot## Warning: Removed 1 rows containing missing values (`geom_point()`).## Warning: Removed 1 row containing missing values (`geom_line()`).
### Affect Instability (MSSD)
# initialize MSSD matrix
ema_mssd <- matrix(0, 32, 7)
colnames(ema_mssd) <- c("group", "posbaseline", "possession4", "pospost", "negbaseline", "negsession4", "negpost")
rownames(ema_mssd) <- id_names
for(i in 1:32){
ema_mssd[i,1] <- id_group[i]
}#fill positive mssd
## baseline
mssd <- mssd(betweenday_baseline$positive_mean, group = betweenday_baseline$id, lag = 1, na.rm = TRUE)
for(i in 1:32){
ema_mssd[i,2] <- mssd[i]
}
## session4
mssd <- mssd(betweenday_session4$positive_mean, group = betweenday_session4$id, lag = 1, na.rm = TRUE)
for(i in 1:32){
ema_mssd[i,3] <- mssd[i]
}
## post
mssd <- mssd(betweenday_post$positive_mean, group = betweenday_post$id, lag = 1, na.rm = TRUE)
for(i in 1:32){
ema_mssd[i,4] <- mssd[i]
}#fill negative mssd
## baseline
mssd <- mssd(betweenday_baseline$negative_mean, group = betweenday_baseline$id, lag = 1, na.rm = TRUE)
for(i in 1:32){
ema_mssd[i,5] <- mssd[i]
}
## session 4
mssd <- mssd(betweenday_session4$negative_mean, group = betweenday_session4$id, lag = 1, na.rm = TRUE)
for(i in 1:32){
ema_mssd[i,6] <- mssd[i]
}
## post
mssd <- mssd(betweenday_post$negative_mean, group = betweenday_post$id, lag = 1, na.rm = TRUE)
for(i in 1:32){
ema_mssd[i,7] <- mssd[i]
}ema_mssd <- as.data.frame(ema_mssd)
ema_mssd$posbaseline <- as.numeric(ema_mssd$posbaseline)
ema_mssd$possession4 <- as.numeric(ema_mssd$possession4)
ema_mssd$pospost <- as.numeric(ema_mssd$pospost)
ema_mssd$negbaseline <- as.numeric(ema_mssd$negbaseline)
ema_mssd$negsession4 <- as.numeric(ema_mssd$negsession4)
ema_mssd$negpost <- as.numeric(ema_mssd$negpost)#group by group
ema_mssd%>%
group_by(group) %>%
summarise_at(vars(posbaseline, possession4, pospost, negbaseline, negsession4, negpost, ), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 2 × 13
## group posbas…¹ posse…² pospo…³ negba…⁴ negse…⁵ negpo…⁶ posba…⁷ posse…⁸ pospo…⁹
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0 18.0 28.0 16.2 15.2 8.34 14.2 17.1 27.2 22.6
## 2 1 27.2 31.0 14.9 15.6 9.80 7.84 38.9 40.8 11.3
## # … with 3 more variables: negbaseline_sd <dbl>, negsession4_sd <dbl>,
## # negpost_sd <dbl>, and abbreviated variable names ¹posbaseline_mean,
## # ²possession4_mean, ³pospost_mean, ⁴negbaseline_mean, ⁵negsession4_mean,
## # ⁶negpost_mean, ⁷posbaseline_sd, ⁸possession4_sd, ⁹pospost_sdPlots
#create id variable
ema_mssd$id <- id_names
#reformat data: baseline
ema_mssd_baseline <- ema_mssd %>%
dplyr::select(id, group, pos = posbaseline, neg = negbaseline)
#create time variable
ema_mssd_baseline$time <- as.integer(1)
ema_mssd_baseline <- relocate(ema_mssd_baseline, time, .before=group)#reformat data: session4
ema_mssd_session4 <- ema_mssd %>%
dplyr::select(id, group, pos = possession4, neg = negsession4)
#create time variable
ema_mssd_session4 $time <- as.integer(2)
ema_mssd_session4 <- relocate(ema_mssd_session4 , time, .before=group)#reformat data: post
ema_mssd_post <- ema_mssd %>%
dplyr::select(id, group, pos = pospost, neg = negpost)
#create time variable
ema_mssd_post$time <- as.integer(3)
ema_mssd_post<- relocate(ema_mssd_post, time, .before=group)#join predata and postdata
ema_mssd2 <- merge(ema_mssd_baseline, ema_mssd_session4, all=TRUE)
ema_mssd2 <- merge(ema_mssd2, ema_mssd_post, all=TRUE)
ema_mssd2 $id <- factor(ema_mssd2$id)
ema_mssd2 $group <- factor(ema_mssd2$group)
levels(ema_mssd2 $group) <- c("Control", "Treatment")
ema_mssd2 $time <- factor(ema_mssd2$time)
levels(ema_mssd2$time) <- c("baseline", "session4", "post")#plot positive means
mssdpos_ggplot<- ggplot(data= ema_mssd2,
aes(y=pos, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("MSSD of Positive Affect Scores Across Timepoints") +
labs(y = "mssd", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
mssdpos_ggplot
#plot negative means
mssdneg_ggplot<- ggplot(data= ema_mssd2,
aes(y=neg, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("MSSD of Negative Affect Scores Across Timepoints") +
labs(y = "mssd", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
mssdneg_ggplot
Acute Increase (PAC)
# initialize PAC matrix
#baseline
ema_pac <- matrix(0, 32, 7)
colnames(ema_pac) <- c("group", "posbaseline", "possession4", "pospost", "negbaseline", "negsession4", "negpost")
rownames(ema_pac) <- id_names
for(i in 1:32){
ema_pac[i,1] <- id_group[i]
}#compute PAC for positive score
pac_vec <- c()
events <- c("Baseline", "Session 4", "Post")
for(n in 1:32){ #cycle through participant ids
name <- id_names[n]
for(e in 1:3){ #cycle through events
index <- which(EMAdata$id == name & EMAdata$event == events[e])
vec <- c()
len <- length(index)
for(i in 1:len){ #number of instances
n <- EMAdata$positive_SD[index[i]]
vec <- c(vec,n) #create vector of pos scores for each event
}
perc_90 <- quantile(vec, 0.9, na.rm=TRUE)
ac <- ifelse(vec >= perc_90, 1, 0)
pac <- sum(ac/(length(ac)-1),na.rm = TRUE)
pac_vec <- c(pac_vec, pac)
}
}pacpos <- matrix(pac_vec, nrow=32)
for(r in 1:32){
ema_pac[r,2] <- pacpos[r,1]
ema_pac[r,3] <- pacpos[r,2]
ema_pac[r,4] <- pacpos[r,3]
}#compute PAC for negative score
pac_vec <- c()
events <- c("Baseline", "Session 4", "Post")
for(n in 1:32){ #cycle through participant ids
name <- id_names[n]
for(e in 1:3){ #cycle through events
index <- which(EMAdata$id == name & EMAdata$event == events[e])
vec <- c()
len <- length(index)
for(i in 1:len){ #number of instances
n <- EMAdata$negative_SD[index[i]]
vec <- c(vec,n) #create vector of pos scores for each event
}
perc_90 <- quantile(vec, 0.9, na.rm=TRUE)
ac <- ifelse(vec >= perc_90, 1, 0)
pac <- sum(ac/(length(ac)-1),na.rm = TRUE)
pac_vec <- c(pac_vec, pac)
}
}pacneg <- matrix(pac_vec, nrow=32)
for(r in 1:32){
ema_pac[r,5] <- pacneg[r,1]
ema_pac[r,6] <- pacneg[r,2]
ema_pac[r,7] <- pacneg[r,3]
}ema_pac <- as.data.frame(ema_pac)
ema_pac$posbaseline <- as.numeric(ema_pac$posbaseline)
ema_pac$possession4 <- as.numeric(ema_pac$possession4)
ema_pac$pospost <- as.numeric(ema_pac$pospost)
ema_pac$negbaseline <- as.numeric(ema_pac$negbaseline)
ema_pac$negsession4 <- as.numeric(ema_pac$negsession4)
ema_pac$negpost <- as.numeric(ema_pac$negpost)#group by group
ema_pac%>%
group_by(group) %>%
summarise_at(vars(posbaseline, possession4, pospost, negbaseline, negsession4, negpost, ), list(mean=mean, sd=sd), na.rm=TRUE)## # A tibble: 2 × 13
## group posbas…¹ posse…² pospo…³ negba…⁴ negse…⁵ negpo…⁶ posba…⁷ posse…⁸ pospo…⁹
## <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
## 1 0 0.171 0.143 0.157 0.181 0.167 0.186 0.0591 0.0468 0.0400
## 2 1 0.172 0.160 0.181 0.168 0.155 0.164 0.0509 0.0312 0.0625
## # … with 3 more variables: negbaseline_sd <dbl>, negsession4_sd <dbl>,
## # negpost_sd <dbl>, and abbreviated variable names ¹posbaseline_mean,
## # ²possession4_mean, ³pospost_mean, ⁴negbaseline_mean, ⁵negsession4_mean,
## # ⁶negpost_mean, ⁷posbaseline_sd, ⁸possession4_sd, ⁹pospost_sdPlots
#create id variable
ema_pac$id <- id_names
#reformat data: baseline
ema_pac_baseline <- ema_pac %>%
dplyr::select(id, group, pos = posbaseline, neg = negbaseline)
#create time variable
ema_pac_baseline$time <- as.integer(1)
ema_pac_baseline <- relocate(ema_pac_baseline, time, .before=group)#reformat data: session4
ema_pac_session4 <- ema_pac %>%
dplyr::select(id, group, pos = possession4, neg = negsession4)
#create time variable
ema_pac_session4 $time <- as.integer(2)
ema_pac_session4 <- relocate(ema_pac_session4 , time, .before=group)#reformat data: post
ema_pac_post <- ema_pac %>%
dplyr::select(id, group, pos = pospost, neg = negpost)
#create time variable
ema_pac_post$time <- as.integer(3)
ema_pac_post<- relocate(ema_pac_post, time, .before=group)#join predata and postdata
ema_pac2 <- merge(ema_pac_baseline, ema_pac_session4, all=TRUE)
ema_pac2 <- merge(ema_pac2, ema_pac_post, all=TRUE)
ema_pac2 $id <- factor(ema_pac2$id)
ema_pac2 $group <- factor(ema_pac2$group)
levels(ema_pac2 $group) <- c("Control", "Treatment")
ema_pac2 $time <- factor(ema_pac2$time)
levels(ema_pac2$time) <- c("baseline", "session4", "post")#plot positive means
pacpos_ggplot<- ggplot(data= ema_pac2,
aes(y=pos, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("PAC of Positive Affect Scores Across Timepoints") +
labs(y = "mssd", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
pacpos_ggplot
#plot negative means
pacneg_ggplot<- ggplot(data= ema_pac2,
aes(y=neg, x=time, color=group, group=id)) +
geom_point() + geom_line() +
ggtitle("PAC of Negative Affect Scores Across Timepoints") +
labs(y = "mssd", x = "timepoint") +
scale_color_manual(values = c("lightblue3","darkblue"))
pacneg_ggplot
Preliminary Analyses
mixed model w REML fixed effects: enrollment group, event, event*enrollmentgroup random effects: id
calculated in excel - kept running into issues doing it in R
filename <- "20230327_betweenday.csv"
betweenday <- read_csv(filename)## Rows: 286 Columns: 16
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (1): date
## dbl (12): id, event, day, instance, enrollmentgroup, completed, positive_sc...
## time (3): time, time_participant, time_interval
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.x Events
lmm_emapos <- lmer(positive_mean ~ enrollmentgroup * event + (1|id), data=betweenday)
summary(lmm_emapos)## Linear mixed model fit by REML ['lmerMod']
## Formula: positive_mean ~ enrollmentgroup * event + (1 | id)
## Data: betweenday
##
## REML criterion at convergence: 1729.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9201 -0.5531 -0.1101 0.5542 2.6795
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 24.04 4.903
## Residual 20.68 4.548
## Number of obs: 283, groups: id, 33
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 21.9437 1.6064 13.660
## enrollmentgroup -1.2525 2.2292 -0.562
## event 0.6044 0.4806 1.258
## enrollmentgroup:event 0.1940 0.6660 0.291
##
## Correlation of Fixed Effects:
## (Intr) enrllm event
## enrllmntgrp -0.721
## event -0.593 0.428
## enrllmntgr: 0.428 -0.592 -0.722anova(lmm_emapos) ## Analysis of Variance Table
## npar Sum Sq Mean Sq F value
## enrollmentgroup 1 4.872 4.872 0.2356
## event 1 92.984 92.984 4.4961
## enrollmentgroup:event 1 1.755 1.755 0.0849eta_squared(lmm_emapos, partial=FALSE)## Warning: Currently only supports partial eta squared for this class of objects.## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------------------
## enrollmentgroup | 4.44e-03 | [0.00, 1.00]
## event | 6.32e-03 | [0.00, 1.00]
## enrollmentgroup:event | 3.41e-04 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].lmm_emaneg <- lmer(negative_mean ~ enrollmentgroup * event + (1|id), data=betweenday)
summary(lmm_emaneg)## Linear mixed model fit by REML ['lmerMod']
## Formula: negative_mean ~ enrollmentgroup * event + (1 | id)
## Data: betweenday
##
## REML criterion at convergence: 1361.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4655 -0.6459 -0.0633 0.4641 4.3345
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.9872 0.9936
## Residual 6.5709 2.5634
## Number of obs: 283, groups: id, 33
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 16.49769 0.63478 25.990
## enrollmentgroup 1.15568 0.87778 1.317
## event -0.03775 0.27048 -0.140
## enrollmentgroup:event -0.53124 0.37478 -1.417
##
## Correlation of Fixed Effects:
## (Intr) enrllm event
## enrllmntgrp -0.723
## event -0.850 0.615
## enrllmntgr: 0.613 -0.849 -0.722anova(lmm_emaneg) ## Analysis of Variance Table
## npar Sum Sq Mean Sq F value
## enrollmentgroup 1 0.3255 0.3255 0.0495
## event 1 18.5346 18.5346 2.8207
## enrollmentgroup:event 1 13.2022 13.2022 2.0092eta_squared(lmm_emaneg, partial=FALSE)## Warning: Currently only supports partial eta squared for this class of objects.## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## -----------------------------------------------------
## enrollmentgroup | 7.90e-03 | [0.00, 1.00]
## event | 7.79e-05 | [0.00, 1.00]
## enrollmentgroup:event | 7.97e-03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].Plots
# Positive Affect plot
xlabels <- c("Baseline", "Session 4", "Post")
#create trendline
ci_panaspos <- confint(lmm_emapos, method="boot", nsim=10)## Computing bootstrap confidence intervals ...## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpointsci_panaspos <- tibble::rownames_to_column(data.frame(ci_panaspos), "Term")
colnames(ci_panaspos)<- c("Term", "CI 2.5%", "CI 97.5%")
as.tibble(ci_panaspos)## Warning: `as.tibble()` was deprecated in tibble 2.0.0.
## ℹ Please use `as_tibble()` instead.
## ℹ The signature and semantics have changed, see `?as_tibble`.## # A tibble: 6 × 3
## Term `CI 2.5%` `CI 97.5%`
## <chr> <dbl> <dbl>
## 1 .sig01 4.41 5.81
## 2 .sigma 4.27 4.77
## 3 (Intercept) 19.5 26.1
## 4 enrollmentgroup -8.90 2.07
## 5 event 0.254 1.23
## 6 enrollmentgroup:event -0.528 0.962#Plot model - positive
panaspos_ggplot<- ggplot(data= betweenday,
aes(y=positive_mean, x=event, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Positive Affect Across Timepoints") +
labs(y = "positive affect score", x = "timepoint") +
theme(legend.position = "none") +
scale_x_discrete(labels = xlabels) +
geom_abline(intercept = fixef(lmm_emapos)[1], #Regression Line (RL).
slope=fixef(lmm_emapos)[2], col="black", linetype = "dashed") +
geom_abline(intercept = ci_panaspos$`CI 97.5%`[3],
slope=ci_panaspos$`CI 97.5%`[4], col="grey", linetype = "dotted") + #Upper Bound of RL
geom_abline(intercept = ci_panaspos$`CI 2.5%`[3],
slope=ci_panaspos$`CI 2.5%`[4], col="grey", linetype= "dotted") #Lower Bound of RL
panaspos_ggplot## Warning: Removed 3 rows containing missing values (`geom_point()`).## Warning: Removed 3 rows containing missing values (`geom_line()`).
# Negative Affect plot
xlabels <- c("Baseline", "Session 4", "Post")
#create trendline
ci_panasneg <- confint(lmm_emaneg, method="boot", nsim=10)## Computing bootstrap confidence intervals ...## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpoints
## Warning in norm.inter(t, alpha): extreme order statistics used as endpointsci_panasneg <- tibble::rownames_to_column(data.frame(ci_panasneg), "Term")
colnames(ci_panasneg)<- c("Term", "CI 2.5%", "CI 97.5%")
as.tibble(ci_panasneg)## # A tibble: 6 × 3
## Term `CI 2.5%` `CI 97.5%`
## <chr> <dbl> <dbl>
## 1 .sig01 0.479 1.46
## 2 .sigma 2.50 2.74
## 3 (Intercept) 15.1 17.4
## 4 enrollmentgroup -0.0586 2.75
## 5 event -0.340 0.404
## 6 enrollmentgroup:event -1.11 -0.0456#Plot model
panasneg_ggplot<- ggplot(data= betweenday,
aes(y=negative_mean, x=event, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Negative Affect Across Timepoints") +
labs(y = "negative affect score", x = "timepoint") +
theme(legend.position = "none") +
scale_x_discrete(labels = xlabels) +
geom_abline(intercept = fixef(lmm_emaneg)[1], #Regression Line (RL).
slope=fixef(lmm_emaneg)[2], col="black", linetype = "dashed") +
geom_abline(intercept = ci_panasneg$`CI 97.5%`[3],
slope=ci_panasneg$`CI 97.5%`[4], col="grey", linetype = "dotted") + #Upper Bound of RL
geom_abline(intercept = ci_panasneg$`CI 2.5%`[3],
slope=ci_panasneg$`CI 2.5%`[4], col="grey", linetype= "dotted") #Lower Bound of RL
panasneg_ggplot## Warning: Removed 3 rows containing missing values (`geom_point()`).## Warning: Removed 3 rows containing missing values (`geom_line()`).
x Days
betweenday$day <- as.factor(betweenday$day)
levels(betweenday$day) <- c("Day 1", "Day 2", "Day 3", "Day 4", "Day 5", "Day 6", "Day 7", "Day 8", "Day 9")lmm_emapos <- lmer(positive_mean ~ enrollmentgroup * day + (1|id), data=betweenday)
anova(lmm_emapos) ## Analysis of Variance Table
## npar Sum Sq Mean Sq F value
## enrollmentgroup 1 4.865 4.8647 0.2369
## day 8 236.891 29.6114 1.4420
## enrollmentgroup:day 8 185.041 23.1302 1.1264summary(lmm_emapos)## Linear mixed model fit by REML ['lmerMod']
## Formula: positive_mean ~ enrollmentgroup * day + (1 | id)
## Data: betweenday
##
## REML criterion at convergence: 1676
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.53516 -0.57192 -0.01372 0.53747 2.70835
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 23.92 4.891
## Residual 20.53 4.531
## Number of obs: 283, groups: id, 33
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 22.8416 1.6514 13.831
## enrollmentgroup 0.7329 2.3113 0.317
## dayDay 2 -0.8544 1.6393 -0.521
## dayDay 3 -0.6758 1.6722 -0.404
## dayDay 4 1.2023 1.6393 0.733
## dayDay 5 0.9223 1.6393 0.563
## dayDay 6 0.2023 1.6393 0.123
## dayDay 7 -0.2288 1.6393 -0.140
## dayDay 8 0.9823 1.6393 0.599
## dayDay 9 1.2245 1.6393 0.747
## enrollmentgroup:dayDay 2 -1.1662 2.2590 -0.516
## enrollmentgroup:dayDay 3 -1.4060 2.3023 -0.611
## enrollmentgroup:dayDay 4 -3.3680 2.2590 -1.491
## enrollmentgroup:dayDay 5 -5.0058 2.2776 -2.198
## enrollmentgroup:dayDay 6 -2.3069 2.2776 -1.013
## enrollmentgroup:dayDay 7 -0.3644 2.2785 -0.160
## enrollmentgroup:dayDay 8 0.4162 2.2785 0.183
## enrollmentgroup:dayDay 9 -1.1573 2.2785 -0.508##
## Correlation matrix not shown by default, as p = 18 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need iteta_squared(lmm_emapos, partial = FALSE)## Warning: Currently only supports partial eta squared for this class of objects.## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## ---------------------------------------------------
## enrollmentgroup | 7.43e-03 | [0.00, 1.00]
## day | 0.02 | [0.00, 1.00]
## enrollmentgroup:day | 0.04 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].lmm_emaneg <- lmer(negative_mean ~ enrollmentgroup * day + (1|id), data=betweenday)
anova(lmm_emaneg) ## Analysis of Variance Table
## npar Sum Sq Mean Sq F value
## enrollmentgroup 1 0.330 0.3295 0.0503
## day 8 75.042 9.3803 1.4310
## enrollmentgroup:day 8 54.495 6.8118 1.0392summary(lmm_emaneg)## Linear mixed model fit by REML ['lmerMod']
## Formula: negative_mean ~ enrollmentgroup * day + (1 | id)
## Data: betweenday
##
## REML criterion at convergence: 1325
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.3481 -0.6159 -0.0443 0.4982 4.5235
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 0.9708 0.9853
## Residual 6.5552 2.5603
## Number of obs: 283, groups: id, 33
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 17.0253 0.6699 25.416
## enrollmentgroup -0.1939 0.9442 -0.205
## dayDay 2 -0.4528 0.9131 -0.496
## dayDay 3 -0.6556 0.9312 -0.704
## dayDay 4 -0.7484 0.9131 -0.820
## dayDay 5 -1.2128 0.9131 -1.328
## dayDay 6 -1.2306 0.9131 -1.348
## dayDay 7 -0.3839 0.9131 -0.420
## dayDay 8 -0.7661 0.9131 -0.839
## dayDay 9 -0.1261 0.9131 -0.138
## enrollmentgroup:dayDay 2 0.3254 1.2669 0.257
## enrollmentgroup:dayDay 3 2.2505 1.2902 1.744
## enrollmentgroup:dayDay 4 0.5052 1.2669 0.399
## enrollmentgroup:dayDay 5 0.6789 1.2770 0.532
## enrollmentgroup:dayDay 6 -0.2686 1.2770 -0.210
## enrollmentgroup:dayDay 7 0.3049 1.2772 0.239
## enrollmentgroup:dayDay 8 0.3694 1.2772 0.289
## enrollmentgroup:dayDay 9 -1.3633 1.2772 -1.067##
## Correlation matrix not shown by default, as p = 18 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need iteta_squared(lmm_emaneg, partial = FALSE)## Warning: Currently only supports partial eta squared for this class of objects.## # Effect Size for ANOVA (Type III)
##
## Parameter | Eta2 (partial) | 95% CI
## ---------------------------------------------------
## enrollmentgroup | 2.07e-03 | [0.00, 1.00]
## day | 0.01 | [0.00, 1.00]
## enrollmentgroup:day | 0.03 | [0.00, 1.00]
##
## - One-sided CIs: upper bound fixed at [1.00].Plots
# Positive Affect plot
xlabels <- c("Day 1", "Day 2", "Day 3", "Day 4", "Day 5", "Day 6", "Day 7", "Day 8", "Day 9")
#Plot model - positive
panaspos_ggplot<- ggplot(data= betweenday,
aes(y=positive_mean, x=day, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Positive Affect Across Days") +
labs(y = "positive affect score", x = "day") +
theme(legend.position = "none") +
scale_x_discrete(labels = xlabels)
panaspos_ggplot## Warning: Removed 3 rows containing missing values (`geom_point()`).## Warning: Removed 3 rows containing missing values (`geom_line()`).
# Negative Affect plot
xlabels <- c("Day 1", "Day 2", "Day 3", "Day 4", "Day 5", "Day 6", "Day 7", "Day 8", "Day 9")
#Plot model - positive
panaspos_ggplot<- ggplot(data= betweenday,
aes(y=negative_mean, x=day, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Negative Affect Across Days") +
labs(y = "negative affect score", x = "day") +
theme(legend.position = "none") +
scale_x_discrete(labels = xlabels)
panaspos_ggplot## Warning: Removed 3 rows containing missing values (`geom_point()`).## Warning: Removed 3 rows containing missing values (`geom_line()`).
x Instances
EMAdata$instance <- as.factor(EMAdata$instance)
levels(EMAdata$instance) <- c("1","2","3","4","5","6","7","8","9","10","11","12","13","14","15")lmm_emapos <- lmer(positive_score ~ enrollmentgroup * instance + (1|id), data= EMAdata)
anova(lmm_emapos) ## Analysis of Variance Table
## npar Sum Sq Mean Sq F value
## enrollmentgroup 1 0.89 0.887 0.0268
## instance 14 2986.95 213.353 6.4483
## enrollmentgroup:instance 14 473.82 33.845 1.0229summary(lmm_emapos)## Linear mixed model fit by REML ['lmerMod']
## Formula: positive_score ~ enrollmentgroup * instance + (1 | id)
## Data: EMAdata
##
## REML criterion at convergence: 7896.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6314 -0.5667 -0.0795 0.5492 4.7119
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 27.08 5.204
## Residual 33.09 5.752
## Number of obs: 1242, groups: id, 33
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 24.96158 1.62308 15.379
## enrollmentgroupImmediate Enrollment -3.06792 2.22861 -1.377
## instance2 0.03676 1.29416 0.028
## instance3 -1.31717 1.32381 -0.995
## instance4 -2.50545 1.29935 -1.928
## instance5 -5.40576 1.30753 -4.134
## instance6 -2.42108 1.30664 -1.853
## instance7 0.46633 1.31510 0.355
## instance8 -1.14320 1.29980 -0.880
## instance9 -2.95746 1.28561 -2.300
## instance10 -3.91047 1.34222 -2.913
## instance11 -1.43807 1.32299 -1.087
## instance12 -1.27751 1.31507 -0.971
## instance13 -1.30025 1.29934 -1.001
## instance14 -1.55438 1.31513 -1.182
## instance15 -5.22615 1.31547 -3.973
## enrollmentgroupImmediate Enrollment:instance2 2.71039 1.78096 1.522
## enrollmentgroupImmediate Enrollment:instance3 3.26754 1.81916 1.796
## enrollmentgroupImmediate Enrollment:instance4 4.25859 1.80876 2.354
## enrollmentgroupImmediate Enrollment:instance5 4.00111 1.81513 2.204
## enrollmentgroupImmediate Enrollment:instance6 2.63457 1.77656 1.483
## enrollmentgroupImmediate Enrollment:instance7 2.57019 1.79726 1.430
## enrollmentgroupImmediate Enrollment:instance8 1.39434 1.79168 0.778
## enrollmentgroupImmediate Enrollment:instance9 3.87343 1.80248 2.149
## enrollmentgroupImmediate Enrollment:instance10 1.74580 1.82771 0.955
## enrollmentgroupImmediate Enrollment:instance11 0.99683 1.80811 0.551
## enrollmentgroupImmediate Enrollment:instance12 4.11750 1.81197 2.272
## enrollmentgroupImmediate Enrollment:instance13 3.99321 1.78039 2.243
## enrollmentgroupImmediate Enrollment:instance14 2.85621 1.81530 1.573
## enrollmentgroupImmediate Enrollment:instance15 3.76380 1.82505 2.062##
## Correlation matrix not shown by default, as p = 30 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itlme_emaneg <- lmer(negative_score ~ enrollmentgroup * instance + (1|id), data= EMAdata)
anova(lme_emaneg) ## Analysis of Variance Table
## npar Sum Sq Mean Sq F value
## enrollmentgroup 1 1.20 1.197 0.0942
## instance 14 601.54 42.967 3.3811
## enrollmentgroup:instance 14 90.62 6.473 0.5093summary(lme_emaneg)## Linear mixed model fit by REML ['lmerMod']
## Formula: negative_score ~ enrollmentgroup * instance + (1 | id)
## Data: EMAdata
##
## REML criterion at convergence: 6730.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6836 -0.5522 -0.1338 0.3839 6.0894
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.447 2.906
## Residual 12.708 3.565
## Number of obs: 1242, groups: id, 33
##
## Fixed effects:
## Estimate Std. Error t value
## (Intercept) 16.96455 0.94258 17.998
## enrollmentgroupImmediate Enrollment -0.27901 1.29204 -0.216
## instance2 0.37199 0.80162 0.464
## instance3 -0.50098 0.82041 -0.611
## instance4 -0.73212 0.80526 -0.909
## instance5 -1.40252 0.81033 -1.731
## instance6 -0.63540 0.80978 -0.785
## instance7 -0.41365 0.81502 -0.508
## instance8 -0.45366 0.80554 -0.563
## instance9 -1.01297 0.79675 -1.271
## instance10 -1.79199 0.83182 -2.154
## instance11 0.01560 0.81991 0.019
## instance12 -0.33090 0.81500 -0.406
## instance13 -1.66764 0.80525 -2.071
## instance14 -1.32012 0.81504 -1.620
## instance15 -1.86610 0.81525 -2.289
## enrollmentgroupImmediate Enrollment:instance2 0.04248 1.10341 0.039
## enrollmentgroupImmediate Enrollment:instance3 -1.18564 1.12738 -1.052
## enrollmentgroupImmediate Enrollment:instance4 -0.58221 1.12092 -0.519
## enrollmentgroupImmediate Enrollment:instance5 -0.53164 1.12486 -0.473
## enrollmentgroupImmediate Enrollment:instance6 -0.25795 1.10099 -0.234
## enrollmentgroupImmediate Enrollment:instance7 0.47638 1.11381 0.428
## enrollmentgroupImmediate Enrollment:instance8 -0.36636 1.11034 -0.330
## enrollmentgroupImmediate Enrollment:instance9 0.60958 1.11705 0.546
## enrollmentgroupImmediate Enrollment:instance10 0.35344 1.13267 0.312
## enrollmentgroupImmediate Enrollment:instance11 0.02329 1.12053 0.021
## enrollmentgroupImmediate Enrollment:instance12 0.49568 1.12293 0.441
## enrollmentgroupImmediate Enrollment:instance13 0.98644 1.10336 0.894
## enrollmentgroupImmediate Enrollment:instance14 -0.35677 1.12497 -0.317
## enrollmentgroupImmediate Enrollment:instance15 -0.36472 1.13103 -0.322##
## Correlation matrix not shown by default, as p = 30 > 12.
## Use print(x, correlation=TRUE) or
## vcov(x) if you need itPlots
# Positive Affect plot @ Baseline
EMA_baseline <- filter(EMAdata, event == "Baseline")
#Plot model - positive
instance_pos<- ggplot(data= EMA_baseline,
aes(y=positive_score, x=instance, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Positive Affect Across Instance @ Baseline") +
labs(y = "positive affect score", x = "instance") +
theme(legend.position = "none")
instance_pos## Warning: Removed 82 rows containing missing values (`geom_point()`).## Warning: Removed 24 rows containing missing values (`geom_line()`).
# Positive Affect plot @ Session 4
EMA_session4 <- filter(EMAdata, event == "Session 4")
#Plot model - positive
instance_pos<- ggplot(data= EMA_session4,
aes(y=positive_score, x=instance, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Positive Affect Across Instance @ Session4") +
labs(y = "positive affect score", x = "instance") +
theme(legend.position = "none")
instance_pos## Warning: Removed 44 rows containing missing values (`geom_point()`).## Warning: Removed 16 rows containing missing values (`geom_line()`).
# Positive Affect plot @ Post
EMA_post <- filter(EMAdata, event == "Post")
#Plot model - positive
instance_pos<- ggplot(data= EMA_post,
aes(y=positive_score, x=instance, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Positive Affect Across Instance @ Post") +
labs(y = "positive affect score", x = "instance") +
theme(legend.position = "none")
instance_pos## Warning: Removed 59 rows containing missing values (`geom_point()`).## Warning: Removed 17 rows containing missing values (`geom_line()`).
# Negative Affect plot @ Baseline
instance_neg<- ggplot(data= EMA_baseline,
aes(y=negative_score, x=instance, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Negative Affect Across Instance @ Baseline") +
labs(y = "negative affect score", x = "instance") +
theme(legend.position = "none")
instance_neg## Warning: Removed 82 rows containing missing values (`geom_point()`).## Warning: Removed 24 rows containing missing values (`geom_line()`).
# Negative Affect plot @ Session 4
instance_neg<- ggplot(data= EMA_session4,
aes(y=negative_score, x=instance, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Negative Affect Across Instance @ Session 4") +
labs(y = "negative affect score", x = "instance") +
theme(legend.position = "none")
instance_neg## Warning: Removed 44 rows containing missing values (`geom_point()`).## Warning: Removed 16 rows containing missing values (`geom_line()`).
# Negative Affect plot @ Post
instance_neg<- ggplot(data= EMA_post,
aes(y=negative_score, x=instance, color=id, group=id)) +
geom_point() + geom_line() +
ggtitle("Negative Affect Across Instance @ Post") +
labs(y = "negative affect score", x = "instance") +
theme(legend.position = "none")
instance_neg## Warning: Removed 59 rows containing missing values (`geom_point()`).## Warning: Removed 17 rows containing missing values (`geom_line()`).
Multilevel Model
#combine dataset
mlm <- ema_mean2[,1:3]
mlm$pos_mean <- ema_mean2[,4]
mlm$neg_mean <- ema_mean2[,5]
mlm$pos_sd <- ema_sd2[,4]
mlm$neg_sd <- ema_sd2[,5]
mlm$pos_ac <- ema_ac2[,4]
mlm$neg_ac <- ema_ac2[,5]
mlm$pos_mssd <- ema_mssd2[,4]
mlm$neg_mssd <- ema_mssd2[,5]
mlm$pos_pac <- ema_pac2[,4]
mlm$neg_pac <- ema_pac2[,5]outcome <- read_csv("20230325_phase2data.csv")## Rows: 108 Columns: 9
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (1): event
## dbl (8): id, group, ryff, pss, bdi, bai, fmqi, cerq
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.outcome <- filter(outcome, group != 2)
outcome <- filter(outcome, id != "12046")
outcome <- filter(outcome, event == "baseline") #baseline only
outcome <- outcome[rep(seq_len(nrow(outcome)), each = 3), ]
class(outcome)## [1] "tbl_df" "tbl" "data.frame"outc <- as.data.frame(outcome[,4:8])
class(outc)## [1] "data.frame"mlm$ryff <- (outc[,1])
mlm$pss <- outc[,2]
mlm$bdi <- outc[,3]
mlm$bai <- outc[,4]
mlm$mi <- outc[,5]
view(mlm)
write.csv(mlm, "20230411_mlm.csv")mlm <- read.csv("MLM_USETHIS.csv")library(lmerTest) #tests on lmer objects##
## Attaching package: 'lmerTest'## The following object is masked from 'package:lme4':
##
## lmer## The following object is masked from 'package:stats':
##
## steplibrary(performance) # ICC calculations
library(texreg)## Version: 1.38.6
## Date: 2022-04-06
## Author: Philip Leifeld (University of Essex)
##
## Consider submitting praise using the praise or praise_interactive functions.
## Please cite the JSS article in your publications -- see citation("texreg").##
## Attaching package: 'texreg'## The following object is masked from 'package:tidyr':
##
## extract#library(interactions) # plotting interactionsNull Model
#Random-Intercept Only
nullmodel_pos <- lmer(positive_score ~ 1 + (1|id), data = mlm)
performance::icc(nullmodel_pos)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.433
## Unadjusted ICC: 0.433#Random-Intercept Only
nullmodel_neg <- lmer(negative_score ~ 1 + (1|id), data = mlm)
performance::icc(nullmodel_neg)## # Intraclass Correlation Coefficient
##
## Adjusted ICC: 0.393
## Unadjusted ICC: 0.393Level-1 (Instance)
Fixed Effects
#L1 Fixed Effects
l1model_pos <- lmer(positive_score ~ 1 + instance + (1|id), data = mlm)
summary(l1model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 8040.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.6373 -0.5897 -0.0500 0.6045 4.2817
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 26.72 5.169
## Residual 35.09 5.924
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 23.41100 0.98062 38.17234 23.874 <2e-16 ***
## instance -0.08331 0.03917 1207.66345 -2.127 0.0336 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## instance -0.316anova(nullmodel_pos, l1model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## nullmodel_pos: positive_score ~ 1 + (1 | id)
## l1model_pos: positive_score ~ 1 + instance + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## nullmodel_pos 3 8047.9 8063.2 -4020.9 8041.9
## l1model_pos 4 8045.3 8065.8 -4018.7 8037.3 4.5243 1 0.03342 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1#L1 Fixed Effects
l1model_neg <- lmer(negative_score ~ 1 + instance + (1|id), data = mlm)
summary(l1model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6798
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0042 -0.5198 -0.1522 0.3655 6.3319
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.393 2.897
## Residual 12.913 3.594
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 16.55811 0.55573 39.68587 29.80 < 2e-16 ***
## instance -0.06867 0.02376 1207.96661 -2.89 0.00392 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## instance -0.338anova(nullmodel_neg, l1model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## nullmodel_neg: negative_score ~ 1 + (1 | id)
## l1model_neg: negative_score ~ 1 + instance + (1 | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## nullmodel_neg 3 6807.2 6822.5 -3400.6 6801.2
## l1model_neg 4 6800.8 6821.3 -3396.4 6792.8 8.3338 1 0.003891 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Random Slopes
l1random_pos <- lmer(positive_score ~ 1 + instance + (1|id) +
(0 + instance|id), data= mlm)
summary(l1random_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + (1 | id) + (0 + instance | id)
## Data: mlm
##
## REML criterion at convergence: 8023.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5319 -0.5766 -0.0284 0.5911 4.3979
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 22.8910 4.7845
## id.1 instance 0.0633 0.2516
## Residual 33.9445 5.8262
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 23.36757 0.91618 34.23047 25.505 <2e-16 ***
## instance -0.07392 0.05921 32.96449 -1.248 0.221
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## instance -0.221anova(l1random_pos, l1model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l1model_pos: positive_score ~ 1 + instance + (1 | id)
## l1random_pos: positive_score ~ 1 + instance + (1 | id) + (0 + instance | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l1model_pos 4 8045.3 8065.8 -4018.7 8037.3
## l1random_pos 5 8031.4 8057.1 -4010.7 8021.4 15.894 1 6.699e-05 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1l1random_neg <- lmer(negative_score ~ 1 + instance + (1|id) +
(0 + instance|id), data= mlm)
summary(l1random_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + (1 | id) + (0 + instance | id)
## Data: mlm
##
## REML criterion at convergence: 6783.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5960 -0.5264 -0.1387 0.3486 6.0853
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.8866 2.9810
## id.1 instance 0.0227 0.1507
## Residual 12.4519 3.5287
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 16.52490 0.56852 33.81751 29.067 <2e-16 ***
## instance -0.06135 0.03563 30.56734 -1.722 0.0952 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr)
## instance -0.217anova(l1random_neg, l1model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l1model_neg: negative_score ~ 1 + instance + (1 | id)
## l1random_neg: negative_score ~ 1 + instance + (1 | id) + (0 + instance | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l1model_neg 4 6800.8 6821.3 -3396.4 6792.8
## l1random_neg 5 6789.1 6814.7 -3389.6 6779.1 13.731 1 0.0002109 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Level-2 (Day)
Fixed Effects
l2model_pos <- lmer(positive_score ~ 1 + instance + day +
(1|id) + (0 + instance|id), data= mlm)
summary(l2model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + (1 | id) + (0 + instance |
## id)
## Data: mlm
##
## REML criterion at convergence: 7994.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5174 -0.5429 -0.0335 0.5527 4.6759
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 22.49770 4.7432
## id.1 instance 0.06287 0.2507
## Residual 33.19207 5.7613
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 21.4594 0.9755 45.4452 21.999 < 2e-16 ***
## instance -0.6651 0.1252 489.7917 -5.313 1.64e-07 ***
## day 3.3056 0.6178 1184.3983 5.350 1.05e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc
## instance 0.226
## day -0.366 -0.883anova(l1random_pos, l2model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l1random_pos: positive_score ~ 1 + instance + (1 | id) + (0 + instance | id)
## l2model_pos: positive_score ~ 1 + instance + day + (1 | id) + (0 + instance | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l1random_pos 5 8031.4 8057.1 -4010.7 8021.4
## l2model_pos 6 8005.1 8035.8 -3996.5 7993.1 28.351 1 1.012e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1l2model_neg <- lmer(negative_score ~ 1 + instance + day +
(1|id) + (0 + instance|id), data= mlm)
summary(l2model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + (1 | id) + (0 + instance |
## id)
## Data: mlm
##
## REML criterion at convergence: 6757.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5588 -0.5383 -0.1254 0.3518 5.9629
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.90385 2.9839
## id.1 instance 0.02291 0.1514
## Residual 12.19419 3.4920
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 15.42320 0.60790 44.10511 25.371 < 2e-16 ***
## instance -0.40255 0.07584 475.56427 -5.308 1.71e-07 ***
## day 1.90816 0.37450 1172.25636 5.095 4.06e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc
## instance 0.220
## day -0.356 -0.883anova(l1random_neg, l2model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l1random_neg: negative_score ~ 1 + instance + (1 | id) + (0 + instance | id)
## l2model_neg: negative_score ~ 1 + instance + day + (1 | id) + (0 + instance | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l1random_neg 5 6789.1 6814.7 -3389.6 6779.1
## l2model_neg 6 6765.4 6796.1 -3376.7 6753.4 25.708 1 3.972e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Random Slopes
l2random_pos <- lmer(positive_score ~ 1 + instance + day +
(1|id) + (0 + instance|id) + (0 + day|id),
data= mlm)## boundary (singular) fit: see help('isSingular')summary(l2random_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + (1 | id) + (0 + instance |
## id) + (0 + day | id)
## Data: mlm
##
## REML criterion at convergence: 7994.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.5174 -0.5429 -0.0335 0.5527 4.6759
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 22.49152 4.7425
## id.1 instance 0.06286 0.2507
## id.2 day 0.00000 0.0000
## Residual 33.19235 5.7613
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 21.4594 0.9754 45.4656 22.001 < 2e-16 ***
## instance -0.6651 0.1252 489.8648 -5.313 1.64e-07 ***
## day 3.3056 0.6179 1184.3939 5.350 1.05e-07 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc
## instance 0.227
## day -0.366 -0.883
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')anova(l2model_pos, l2random_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l2model_pos: positive_score ~ 1 + instance + day + (1 | id) + (0 + instance | id)
## l2random_pos: positive_score ~ 1 + instance + day + (1 | id) + (0 + instance | id) + (0 + day | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l2model_pos 6 8005.1 8035.8 -3996.5 7993.1
## l2random_pos 7 8007.1 8043.0 -3996.5 7993.1 0 1 1l2random_neg <- lmer(negative_score ~ 1 + instance + day +
(1|id) + (0 + instance|id) + (0 + day|id),
data= mlm)
summary(l2random_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + (1 | id) + (0 + instance |
## id) + (0 + day | id)
## Data: mlm
##
## REML criterion at convergence: 6752.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.4727 -0.5293 -0.1162 0.3489 6.0696
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.724935 2.95380
## id.1 instance 0.005796 0.07613
## id.2 day 0.606159 0.77856
## Residual 12.088054 3.47679
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 15.40876 0.60331 39.90184 25.541 < 2e-16 ***
## instance -0.40666 0.07188 323.87748 -5.658 3.38e-08 ***
## day 1.92933 0.39827 383.75487 4.844 1.85e-06 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc
## instance 0.235
## day -0.338 -0.870anova(l2model_neg, l2random_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l2model_neg: negative_score ~ 1 + instance + day + (1 | id) + (0 + instance | id)
## l2random_neg: negative_score ~ 1 + instance + day + (1 | id) + (0 + instance | id) + (0 + day | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l2model_neg 6 6765.4 6796.1 -3376.7 6753.4
## l2random_neg 7 6761.9 6797.8 -3374.0 6747.9 5.47 1 0.01935 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Level-3 (Event)
Fixed Effects
l3model_pos <- lmer(positive_score ~ 1 + instance + day + event +
(1|id) + (0 + instance|id), data= mlm)
summary(l3model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + (1 | id) + (0 +
## instance | id)
## Data: mlm
##
## REML criterion at convergence: 7950.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3328 -0.6053 -0.0121 0.5105 4.5273
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 22.50731 4.7442
## id.1 instance 0.06286 0.2507
## Residual 31.98979 5.6560
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 18.7438 1.0508 61.3227 17.838 < 2e-16 ***
## instance -0.6497 0.1232 476.7678 -5.273 2.04e-07 ***
## day 3.2475 0.6066 1182.9940 5.353 1.04e-07 ***
## event 1.3569 0.2007 1179.7832 6.763 2.13e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day
## instance 0.199
## day -0.328 -0.881
## event -0.382 0.018 -0.014anova(l2model_pos, l3model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l2model_pos: positive_score ~ 1 + instance + day + (1 | id) + (0 + instance | id)
## l3model_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l2model_pos 6 8005.1 8035.8 -3996.5 7993.1
## l3model_pos 7 7962.1 7998.0 -3974.1 7948.1 44.973 1 1.998e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1l3model_neg <- lmer(negative_score ~ 1 + instance + day + event +
(1|id) + (0 + instance|id) + (0 + day|id),
data= mlm)
summary(l3model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + (1 | id) + (0 +
## instance | id) + (0 + day | id)
## Data: mlm
##
## REML criterion at convergence: 6751.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.5400 -0.5506 -0.1118 0.3503 6.0143
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.731818 2.95496
## id.1 instance 0.005723 0.07565
## id.2 day 0.607983 0.77973
## Residual 12.065446 3.47354
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 15.85093 0.65180 54.32332 24.319 < 2e-16 ***
## instance -0.40906 0.07181 323.80669 -5.697 2.75e-08 ***
## day 1.93852 0.39805 383.52098 4.870 1.63e-06 ***
## event -0.22096 0.12323 1150.14318 -1.793 0.0732 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day
## instance 0.210
## day -0.307 -0.870
## event -0.378 0.019 -0.013anova(l2random_neg, l3model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l2random_neg: negative_score ~ 1 + instance + day + (1 | id) + (0 + instance | id) + (0 + day | id)
## l3model_neg: negative_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + day | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l2random_neg 7 6761.9 6797.8 -3374.0 6747.9
## l3model_neg 8 6760.7 6801.7 -3372.4 6744.7 3.2159 1 0.07292 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Random Effects
l3random_pos <- lmer(positive_score ~ 1 + instance + day + event +
(1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)
summary(l3random_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + (1 | id) + (0 +
## instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7792.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9869 -0.5841 -0.0094 0.5641 4.0528
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 39.99764 6.3244
## id.1 instance 0.06029 0.2455
## id.2 event 7.76956 2.7874
## Residual 26.09718 5.1085
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 18.9336 1.2602 36.0397 15.025 < 2e-16 ***
## instance -0.6696 0.1128 421.0196 -5.934 6.17e-09 ***
## day 3.3404 0.5494 1137.1313 6.080 1.64e-09 ***
## event 1.2054 0.5288 31.1688 2.280 0.0296 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day
## instance 0.148
## day -0.246 -0.871
## event -0.108 0.007 -0.007anova(l3model_pos, l3random_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3model_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id)
## l3random_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3model_pos 7 7962.1 7998.0 -3974.1 7948.1
## l3random_pos 8 7807.6 7848.6 -3895.8 7791.6 156.51 1 < 2.2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1l3random_neg <- lmer(negative_score ~ 1 + instance + day + event +
(1|id) + (0 + instance|id) + (0 + day|id) +
(0 + event|id), data= mlm)
summary(l3random_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + (1 | id) + (0 +
## instance | id) + (0 + day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6692.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7614 -0.5512 -0.1165 0.3426 5.8309
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.887714 2.98123
## id.1 instance 0.008473 0.09205
## id.2 day 0.852399 0.92325
## id.3 event 1.430644 1.19610
## Residual 10.951782 3.30935
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 16.01054 0.64699 38.08214 24.746 < 2e-16 ***
## instance -0.41459 0.06935 253.77099 -5.978 7.63e-09 ***
## day 1.97995 0.39254 291.02557 5.044 8.04e-07 ***
## event -0.33384 0.24423 30.00279 -1.367 0.182
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day
## instance 0.199
## day -0.285 -0.834
## event -0.185 0.012 -0.010anova(l3model_neg, l3random_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3model_neg: negative_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + day | id)
## l3random_neg: negative_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3model_neg 8 6760.7 6801.7 -3372.4 6744.7
## l3random_neg 9 6705.7 6751.8 -3343.9 6687.7 56.989 1 4.383e-14 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Level-4 (Participant)
Fixed Effects
l4model_pos <- lmer(positive_score ~ 1 + instance + day + event + group +
(1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)
summary(l4model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + group + (1 | id) +
## (0 + instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7788.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9870 -0.5836 -0.0093 0.5637 4.0520
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 41.4187 6.4357
## id.1 instance 0.0607 0.2464
## id.2 event 7.8084 2.7944
## Residual 26.0912 5.1080
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 19.2834 1.8136 30.8734 10.633 7.63e-12 ***
## instance -0.6696 0.1129 418.3341 -5.932 6.29e-09 ***
## day 3.3406 0.5494 1137.6212 6.081 1.63e-09 ***
## event 1.2056 0.5300 31.1639 2.275 0.0299 *
## group -0.6591 2.4440 28.5776 -0.270 0.7894
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event
## instance 0.101
## day -0.168 -0.871
## event -0.068 0.007 -0.007
## group -0.710 0.002 -0.005 -0.009anova(l3random_pos, l4model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + event | id)
## l4model_pos: positive_score ~ 1 + instance + day + event + group + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_pos 8 7807.6 7848.6 -3895.8 7791.6
## l4model_pos 9 7809.5 7855.6 -3895.8 7791.5 0.0797 1 0.7777l4model_neg <- lmer(negative_score ~ 1 + instance + day + event + group +
(1|id) + (0 + instance|id) + (0 + day|id) +
(0 + event|id), data= mlm)
summary(l4model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + group + (1 | id) +
## (0 + instance | id) + (0 + day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6690.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7640 -0.5471 -0.1162 0.3393 5.8382
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 9.173949 3.02885
## id.1 instance 0.008665 0.09309
## id.2 day 0.865456 0.93030
## id.3 event 1.429830 1.19576
## Residual 10.948115 3.30879
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 16.34097 0.91781 31.17732 17.804 < 2e-16 ***
## instance -0.41471 0.06939 250.25579 -5.976 7.81e-09 ***
## day 1.98172 0.39305 284.93247 5.042 8.21e-07 ***
## event -0.33323 0.24419 29.96636 -1.365 0.183
## group -0.62788 1.22993 28.07380 -0.511 0.614
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event
## instance 0.136
## day -0.193 -0.833
## event -0.123 0.011 -0.010
## group -0.702 0.006 -0.010 -0.010anova(l3random_neg, l4model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_neg: negative_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## l4model_neg: negative_score ~ 1 + instance + day + event + group + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_neg 9 6705.7 6751.8 -3343.9 6687.7
## l4model_neg 10 6707.4 6758.7 -3343.7 6687.4 0.276 1 0.5993Explanatory Variables
Level-2 (Day)
ex1model_pos <- lmer(positive_score ~ 1 + instance + day + event +
weekend + (1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)
summary(ex1model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + weekend + (1 |
## id) + (0 + instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7792.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9861 -0.5840 -0.0089 0.5645 4.0499
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 40.05787 6.3291
## id.1 instance 0.06107 0.2471
## id.2 event 7.77242 2.7879
## Residual 26.11287 5.1101
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 18.94303 1.27796 37.87909 14.823 < 2e-16 ***
## instance -0.66948 0.11298 413.86975 -5.926 6.56e-09 ***
## day 3.33842 0.55108 1138.76146 6.058 1.87e-09 ***
## event 1.20518 0.52889 31.15316 2.279 0.0297 *
## weekend -0.01779 0.38592 319.22112 -0.046 0.9633
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event
## instance 0.147
## day -0.254 -0.869
## event -0.108 0.007 -0.006
## weekend -0.162 -0.010 0.073 0.007anova(l3random_pos, ex1model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + event | id)
## ex1model_pos: positive_score ~ 1 + instance + day + event + weekend + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_pos 8 7807.6 7848.6 -3895.8 7791.6
## ex1model_pos 9 7809.6 7855.7 -3895.8 7791.6 0.003 1 0.9564ex1model_neg <- lmer(negative_score ~ 1 + instance + day + event +
weekend +(1|id) + (0 + instance|id) + (0 + day|id) +
(0 + event|id), data= mlm)
summary(ex1model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + weekend + (1 |
## id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6693.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7687 -0.5528 -0.1157 0.3327 5.8376
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.928575 2.9881
## id.1 instance 0.008537 0.0924
## id.2 day 0.857085 0.9258
## id.3 event 1.431570 1.1965
## Residual 10.958011 3.3103
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 15.97360 0.66272 41.27582 24.103 < 2e-16 ***
## instance -0.41469 0.06939 251.73339 -5.976 7.77e-09 ***
## day 1.98658 0.39361 291.60959 5.047 7.91e-07 ***
## event -0.33327 0.24432 30.04709 -1.364 0.183
## weekend 0.06999 0.26186 336.80927 0.267 0.789
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event
## instance 0.195
## day -0.291 -0.832
## event -0.182 0.012 -0.009
## weekend -0.209 -0.006 0.063 0.010anova(l3random_neg, ex1model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_neg: negative_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## ex1model_neg: negative_score ~ 1 + instance + day + event + weekend + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_neg 9 6705.7 6751.8 -3343.9 6687.7
## ex1model_neg 10 6707.6 6758.9 -3343.8 6687.6 0.0785 1 0.7794Level-3 (Event)
ex2model_pos <- lmer(positive_score ~ 1 + instance + day + event +
pos_sd + pos_ac + pos_mssd + pos_pac +
(1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)
summary(ex2model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + pos_sd + pos_ac +
## pos_mssd + pos_pac + (1 | id) + (0 + instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7789.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9518 -0.5914 -0.0123 0.5717 4.0445
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 38.84974 6.2330
## id.1 instance 0.05864 0.2422
## id.2 event 8.17175 2.8586
## Residual 26.08189 5.1070
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.775e+01 1.710e+00 9.730e+01 10.380 < 2e-16 ***
## instance -6.687e-01 1.126e-01 4.317e+02 -5.939 5.90e-09 ***
## day 3.321e+00 5.494e-01 1.129e+03 6.044 2.04e-09 ***
## event 5.413e-01 6.562e-01 5.996e+01 0.825 0.4127
## pos_sd 6.196e-02 3.552e-02 1.192e+03 1.744 0.0814 .
## pos_ac -1.200e+01 7.327e+00 5.411e+02 -1.638 0.1021
## pos_mssd -3.504e-03 9.964e-03 9.860e+02 -0.352 0.7252
## pos_pac -1.674e+00 4.771e+00 1.041e+03 -0.351 0.7257
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event pos_sd pos_ac ps_mss
## instance 0.112
## day -0.176 -0.873
## event -0.087 0.001 0.003
## pos_sd 0.062 0.010 -0.016 -0.564
## pos_ac 0.480 0.002 0.012 0.175 -0.272
## pos_mssd -0.215 0.003 -0.002 0.232 -0.338 0.037
## pos_pac -0.280 -0.003 0.004 0.116 -0.230 0.298 0.099anova(l3random_pos, ex2model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + event | id)
## ex2model_pos: positive_score ~ 1 + instance + day + event + pos_sd + pos_ac + pos_mssd + pos_pac + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_pos 8 7807.6 7848.6 -3895.8 7791.6
## ex2model_pos 12 7811.0 7872.4 -3893.5 7787.0 4.6496 4 0.3252ex2model_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
(1|id) + (0 + instance|id) + (0 + day|id) + (0 + event|id),
data= mlm)
summary(ex2model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + (1 | id) + (0 + instance | id) + (0 +
## day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6657.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7773 -0.5208 -0.1232 0.3182 5.7669
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 7.497391 2.73814
## id.1 instance 0.005555 0.07453
## id.2 day 0.833031 0.91271
## id.3 event 2.053386 1.43296
## Residual 10.604288 3.25642
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.22431 0.95836 140.02603 12.755 < 2e-16 ***
## instance -0.41125 0.06763 277.08814 -6.080 3.96e-09 ***
## day 1.94118 0.38673 319.67575 5.019 8.62e-07 ***
## event -1.12469 0.34894 58.44591 -3.223 0.002075 **
## neg_sd 0.09827 0.02944 1177.66678 3.337 0.000872 ***
## neg_ac -16.96179 4.34401 471.07988 -3.905 0.000108 ***
## neg_mssd 0.03309 0.01136 937.10271 2.913 0.003663 **
## neg_pac 7.03887 2.38696 662.00438 2.949 0.003302 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss
## instance 0.131
## day -0.175 -0.841
## event -0.001 0.000 0.004
## neg_sd -0.045 0.017 -0.019 -0.581
## neg_ac 0.650 -0.008 0.024 0.178 -0.227
## neg_mssd -0.103 0.008 0.001 0.031 0.048 0.129
## neg_pac -0.405 -0.008 0.004 -0.029 -0.128 -0.050 0.036anova(l3random_neg, ex2model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_neg: negative_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## ex2model_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_pac + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_neg 9 6705.7 6751.8 -3343.9 6687.7
## ex2model_neg 13 6674.9 6741.5 -3324.5 6648.9 38.811 4 7.621e-08 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1Level-4 (Participant)
ex3model_pos <- lmer(positive_score ~ 1 + instance + day + event +
ryff + pss + bai + bdi + mi +
(1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)
summary(ex3model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + pss + bai +
## bdi + mi + (1 | id) + (0 + instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7791.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0209 -0.5799 -0.0137 0.5552 4.0907
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 34.9709 5.9136
## id.1 instance 0.0588 0.2425
## id.2 event 8.0657 2.8400
## Residual 26.0818 5.1070
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) -6.96581 15.33743 25.05722 -0.454 0.6536
## instance -0.66824 0.11260 428.16148 -5.934 6.09e-09 ***
## day 3.33047 0.54928 1138.78944 6.063 1.81e-09 ***
## event 1.20171 0.53762 31.29723 2.235 0.0327 *
## ryff 0.31233 0.13075 24.89999 2.389 0.0248 *
## pss -0.12269 0.34385 24.84531 -0.357 0.7242
## bai -0.03522 0.22376 25.54191 -0.157 0.8762
## bdi 0.38251 0.23530 26.23511 1.626 0.1160
## mi 0.01790 0.19746 24.65057 0.091 0.9285
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff pss bai bdi
## instance 0.011
## day -0.016 -0.873
## event -0.003 0.007 -0.007
## ryff -0.743 0.006 -0.009 -0.006
## pss -0.587 -0.003 0.003 0.011 0.183
## bai 0.226 0.006 -0.006 -0.009 0.046 -0.623
## bdi -0.646 -0.004 -0.001 -0.016 0.545 0.214 -0.472
## mi -0.333 -0.004 0.003 -0.004 -0.244 0.128 -0.031 0.081anova(l3random_pos, ex3model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + event | id)
## ex3model_pos: positive_score ~ 1 + instance + day + event + ryff + pss + bai + bdi + mi + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_pos 8 7807.6 7848.6 -3895.8 7791.6
## ex3model_pos 13 7808.1 7874.7 -3891.0 7782.1 9.5108 5 0.09034 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1ex4model_pos <- lmer(positive_score ~ 1 + instance + day + event +
ryff +
(1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.00235948 (tol = 0.002, component 1)summary(ex4model_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + (1 | id) +
## (0 + instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7790
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -4.0043 -0.5816 -0.0085 0.5620 4.0739
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 33.39014 5.7784
## id.1 instance 0.06147 0.2479
## id.2 event 7.85142 2.8020
## Residual 26.08799 5.1076
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 6.34057 5.29808 28.59371 1.197 0.2412
## instance -0.66744 0.11299 413.33960 -5.907 7.27e-09 ***
## day 3.33326 0.54933 1138.44935 6.068 1.76e-09 ***
## event 1.22641 0.53114 31.44420 2.309 0.0277 *
## ryff 0.18307 0.07519 28.04986 2.435 0.0215 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event
## instance 0.027
## day -0.051 -0.870
## event -0.038 0.007 -0.007
## ryff -0.975 0.008 -0.007 0.013
## optimizer (nloptwrap) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.00235948 (tol = 0.002, component 1)anova(l3random_pos, ex4model_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## l3random_pos: positive_score ~ 1 + instance + day + event + (1 | id) + (0 + instance | id) + (0 + event | id)
## ex4model_pos: positive_score ~ 1 + instance + day + event + ryff + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## l3random_pos 8 7807.6 7848.6 -3895.8 7791.6
## ex4model_pos 9 7803.9 7850.0 -3892.9 7785.9 5.7206 1 0.01677 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1ex3model_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + #neg_pac +
ryff + pss + bai + bdi + mi +
(1|id) + (0 + instance|id) + (0 + day|id) + (0 + event|id),
data= mlm)
summary(ex3model_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + ryff + pss + bai + bdi + mi + (1 | id) + (0 +
## instance | id) + (0 + day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6677.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.7088 -0.5414 -0.1273 0.3637 5.7751
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 6.623008 2.57352
## id.1 instance 0.005299 0.07279
## id.2 day 0.685802 0.82813
## id.3 event 1.762343 1.32753
## Residual 10.741627 3.27744
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.578e+01 7.387e+00 2.475e+01 2.136 0.042766 *
## instance -4.093e-01 6.798e-02 3.115e+02 -6.021 4.87e-09 ***
## day 1.928e+00 3.827e-01 3.583e+02 5.037 7.50e-07 ***
## event -1.058e+00 3.362e-01 6.942e+01 -3.148 0.002420 **
## neg_sd 1.024e-01 2.948e-02 1.173e+03 3.473 0.000533 ***
## neg_ac -1.424e+01 4.344e+00 4.680e+02 -3.279 0.001121 **
## neg_mssd 2.863e-02 1.151e-02 9.503e+02 2.488 0.013022 *
## ryff -3.139e-02 6.311e-02 2.480e+01 -0.497 0.623256
## pss -2.186e-01 1.652e-01 2.426e+01 -1.323 0.198127
## bai 2.131e-01 1.083e-01 2.542e+01 1.968 0.060069 .
## bdi -3.275e-03 1.155e-01 2.754e+01 -0.028 0.977576
## mi 6.666e-02 9.480e-02 2.415e+01 0.703 0.488647
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss ryff pss bai
## instance 0.010
## day -0.011 -0.856
## event -0.014 -0.001 0.004
## neg_sd 0.028 0.015 -0.017 -0.613
## neg_ac 0.038 -0.008 0.022 0.182 -0.243
## neg_mssd 0.090 0.007 0.003 0.028 0.067 0.102
## ryff -0.744 0.012 -0.015 0.012 -0.042 0.033 -0.103
## pss -0.577 -0.003 0.003 -0.013 0.041 -0.082 0.042 0.171
## bai 0.223 0.006 -0.005 0.015 -0.042 0.073 -0.059 0.055 -0.629
## bdi -0.647 0.003 -0.012 0.015 -0.063 0.089 -0.123 0.554 0.197 -0.446
## mi -0.331 -0.001 0.000 0.011 -0.027 0.064 -0.044 -0.234 0.116 -0.024
## bdi
## instance
## day
## event
## neg_sd
## neg_ac
## neg_mssd
## ryff
## pss
## bai
## bdi
## mi 0.095anova(ex2model_neg, ex3model_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## ex2model_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_pac + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## ex3model_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + ryff + pss + bai + bdi + mi + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## ex2model_neg 13 6674.9 6741.5 -3324.5 6648.9
## ex3model_neg 17 6682.3 6769.4 -3324.1 6648.3 0.6302 4 0.9597Cross-Level Interactions
Treatment Group
crosslevel1_pos <- lmer(positive_score ~ 1 + instance + day + event + ryff +
group + group:instance + (1|id), data = mlm)
summary(crosslevel1_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + group +
## group:instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 7971.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3442 -0.6189 -0.0472 0.5166 4.3267
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 28.13 5.304
## Residual 33.12 5.755
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 17.86179 4.76478 30.10820 3.749 0.000755 ***
## instance -0.70755 0.12330 1204.73081 -5.739 1.21e-08 ***
## day 3.27851 0.61643 1204.77173 5.319 1.25e-07 ***
## event 1.37024 0.20409 1204.97446 6.714 2.91e-11 ***
## ryff 0.01772 0.06463 28.80623 0.274 0.785932
## group -0.64520 2.00677 34.92727 -0.322 0.749739
## instance:group 0.08330 0.07615 1204.53223 1.094 0.274221
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff group
## instance 0.025
## day -0.074 -0.897
## event -0.087 0.014 -0.014
## ryff -0.945 -0.002 0.002 0.000
## group -0.286 0.104 -0.008 0.002 0.069
## instanc:grp 0.067 -0.324 0.009 0.013 -0.003 -0.301crosslevel1_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
group + group:instance + (1|id), data = mlm)
summary(crosslevel1_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + group + group:instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6752.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9132 -0.5496 -0.1388 0.3736 6.2479
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.28 2.878
## Residual 12.43 3.525
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 14.33722 1.05730 98.63297 13.560 < 2e-16 ***
## instance -0.43059 0.07555 1201.03107 -5.699 1.51e-08 ***
## day 1.88615 0.37778 1201.02009 4.993 6.83e-07 ***
## event -0.83364 0.24788 1213.62927 -3.363 0.000795 ***
## neg_sd 0.09240 0.03090 1217.73177 2.990 0.002847 **
## neg_ac -9.18387 3.53414 880.34531 -2.599 0.009517 **
## neg_mssd 0.03318 0.01074 1229.89580 3.089 0.002052 **
## neg_pac 1.90937 2.01447 1229.94735 0.948 0.343404
## group -0.78503 1.10765 38.51991 -0.709 0.482754
## instance:group 0.04803 0.04665 1201.33440 1.030 0.303450
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc group
## instance 0.053
## day -0.185 -0.897
## event -0.110 -0.011 0.011
## neg_sd 0.017 0.021 -0.021 -0.856
## neg_ac 0.442 -0.027 0.031 0.167 -0.181
## neg_mssd -0.109 0.004 -0.003 0.047 0.021 0.122
## neg_pac -0.329 0.002 0.001 0.011 -0.123 -0.016 0.036
## group -0.534 0.113 -0.007 0.014 -0.018 0.070 0.022 0.042
## instanc:grp 0.181 -0.324 0.009 0.012 -0.004 0.010 0.014 -0.010 -0.333crosslevel2_pos <- lmer(positive_score ~ 1 + instance + day + event + ryff +
group + group:day + (1|id), data = mlm)
summary(crosslevel2_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + group +
## group:day + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 7969
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3320 -0.6206 -0.0507 0.5165 4.3203
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 28.12 5.303
## Residual 33.15 5.757
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 17.69957 4.77148 30.29871 3.709 0.000834 ***
## instance -0.66419 0.11668 1205.03806 -5.692 1.57e-08 ***
## day 3.18260 0.64860 1204.80586 4.907 1.05e-06 ***
## event 1.36842 0.20417 1204.97351 6.702 3.14e-11 ***
## ryff 0.01787 0.06462 28.80890 0.277 0.784102
## group -0.34499 2.07517 39.94493 -0.166 0.868801
## day:group 0.18056 0.40261 1204.29072 0.448 0.653891
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff group
## instance 0.049
## day -0.097 -0.896
## event -0.086 0.019 -0.017
## ryff -0.944 -0.003 0.003 0.000
## group -0.290 0.009 0.115 0.002 0.066
## day:group 0.087 -0.007 -0.310 0.011 -0.002 -0.387crosslevel2_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
group + group:day + (1|id), data = mlm)
summary(crosslevel2_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + group + group:day + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6748.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9048 -0.5426 -0.1360 0.3758 6.2279
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.279 2.877
## Residual 12.421 3.524
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 14.45488 1.07123 103.66376 13.494 < 2e-16 ***
## instance -0.40603 0.07146 1201.19247 -5.682 1.67e-08 ***
## day 1.73275 0.39719 1201.03811 4.363 1.40e-05 ***
## event -0.83417 0.24783 1213.61463 -3.366 0.000787 ***
## neg_sd 0.09249 0.03090 1217.72147 2.993 0.002814 **
## neg_ac -9.12873 3.53417 880.08375 -2.583 0.009955 **
## neg_mssd 0.03323 0.01074 1229.89447 3.094 0.002021 **
## neg_pac 1.90768 2.01411 1229.94818 0.947 0.343743
## group -1.00299 1.15353 45.28943 -0.869 0.389160
## day:group 0.30080 0.24653 1201.19944 1.220 0.222649
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc group
## instance 0.115
## day -0.249 -0.896
## event -0.109 -0.007 0.008
## neg_sd 0.017 0.020 -0.020 -0.856
## neg_ac 0.440 -0.025 0.023 0.167 -0.181
## neg_mssd -0.107 0.009 -0.007 0.047 0.021 0.122
## neg_pac -0.325 -0.002 0.004 0.011 -0.123 -0.016 0.036
## group -0.551 0.009 0.128 0.014 -0.018 0.061 0.019 0.041
## day:group 0.241 -0.008 -0.309 0.008 -0.001 0.021 0.015 -0.009 -0.425crosslevel3_pos <- lmer(positive_score ~ 1 + instance + day + event + ryff +
group + group:event + (1|id), data = mlm)
summary(crosslevel3_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + group +
## group:event + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 7968.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3536 -0.6190 -0.0410 0.5227 4.2892
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 28.12 5.303
## Residual 33.13 5.756
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 17.92750 4.77337 30.34532 3.756 0.000734 ***
## instance -0.66290 0.11665 1205.05041 -5.683 1.66e-08 ***
## day 3.27036 0.61649 1204.78595 5.305 1.34e-07 ***
## event 1.17171 0.28969 1204.31173 4.045 5.57e-05 ***
## ryff 0.01766 0.06462 28.81077 0.273 0.786565
## group -0.76251 2.08063 40.35943 -0.366 0.715920
## event:group 0.38855 0.40814 1204.91144 0.952 0.341290
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff group
## instance 0.050
## day -0.074 -0.945
## event -0.126 0.008 -0.007
## ryff -0.944 -0.003 0.002 0.003
## group -0.292 0.003 -0.004 0.283 0.067
## event:group 0.091 0.008 -0.004 -0.710 -0.004 -0.393crosslevel3_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
group + group:event + (1|id), data = mlm)
summary(crosslevel3_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + group + group:event + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6750.2
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9553 -0.5384 -0.1460 0.3751 6.2792
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.285 2.878
## Residual 12.435 3.526
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 14.22749 1.07464 104.33635 13.239 < 2e-16 ***
## instance -0.40518 0.07150 1201.21063 -5.667 1.82e-08 ***
## day 1.88240 0.37791 1201.02357 4.981 7.25e-07 ***
## event -0.87719 0.27754 1209.18120 -3.161 0.00161 **
## neg_sd 0.09251 0.03092 1217.73546 2.992 0.00283 **
## neg_ac -9.12545 3.54702 877.20701 -2.573 0.01025 *
## neg_mssd 0.03347 0.01083 1229.89869 3.090 0.00204 **
## neg_pac 1.94066 2.01543 1229.95606 0.963 0.33578
## group -0.56596 1.15635 45.66281 -0.489 0.62687
## event:group 0.08194 0.25258 1203.60702 0.324 0.74568
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc group
## instance 0.118
## day -0.184 -0.945
## event -0.212 -0.010 0.011
## neg_sd 0.016 0.020 -0.021 -0.763
## neg_ac 0.453 -0.024 0.030 0.112 -0.180
## neg_mssd -0.078 0.010 -0.003 -0.014 0.021 0.130
## neg_pac -0.318 -0.002 0.001 0.003 -0.123 -0.015 0.037
## group -0.554 0.002 -0.003 0.208 -0.017 0.035 -0.028 0.030
## event:group 0.251 0.008 -0.002 -0.449 -0.004 0.081 0.124 0.016 -0.429Significant Predictors
crosslevel4_pos <- lmer(positive_score ~ 1 + instance + day + event +
ryff + ryff:instance + (1|id), data = mlm)
summary(crosslevel4_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + ryff:instance +
## (1 | id)
## Data: mlm
##
## REML criterion at convergence: 7981.7
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3000 -0.6242 -0.0419 0.5204 4.3543
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 27.20 5.215
## Residual 33.13 5.756
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.634e+01 4.700e+00 3.721e+01 3.476 0.00131 **
## instance -5.087e-01 2.154e-01 1.205e+03 -2.362 0.01834 *
## day 3.260e+00 6.167e-01 1.205e+03 5.286 1.48e-07 ***
## event 1.367e+00 2.041e-01 1.205e+03 6.696 3.27e-11 ***
## ryff 3.516e-02 6.654e-02 3.609e+01 0.528 0.60043
## instance:ryff -2.200e-03 2.566e-03 1.204e+03 -0.857 0.39139
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff
## instance -0.219
## day -0.069 -0.533
## event -0.086 0.007 -0.014
## ryff -0.971 0.252 -0.005 -0.001
## instnc:ryff 0.294 -0.841 0.025 0.004 -0.302crosslevel5_pos <- lmer(positive_score ~ 1 + instance + day + event +
ryff + ryff:day + (1|id), data = mlm)
summary(crosslevel5_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + ryff:day +
## (1 | id)
## Data: mlm
##
## REML criterion at convergence: 7979.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.3157 -0.6185 -0.0478 0.5103 4.3248
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 27.15 5.210
## Residual 33.15 5.758
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.757e+01 4.849e+00 4.227e+01 3.624 0.000772 ***
## instance -6.641e-01 1.167e-01 1.205e+03 -5.690 1.60e-08 ***
## day 3.248e+00 1.107e+00 1.204e+03 2.934 0.003405 **
## event 1.368e+00 2.042e-01 1.205e+03 6.698 3.24e-11 ***
## ryff 1.718e-02 6.875e-02 4.128e+01 0.250 0.803938
## day:ryff 3.776e-04 1.349e-02 1.204e+03 0.028 0.977672
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff
## instance 0.041
## day -0.356 -0.506
## event -0.081 0.019 -0.015
## ryff -0.973 0.006 0.323 -0.003
## day:ryff 0.378 -0.025 -0.830 0.008 -0.388crosslevel6_pos <- lmer(positive_score ~ 1 + instance + day + event +
ryff + ryff:event + (1|id), data = mlm)
summary(crosslevel6_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + ryff:event +
## (1 | id)
## Data: mlm
##
## REML criterion at convergence: 7934.5
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.1740 -0.6409 -0.0446 0.5354 4.0126
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 27.24 5.219
## Residual 31.95 5.652
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.89394 4.86611 42.70454 1.006 0.32022
## instance -0.65463 0.11455 1205.14231 -5.715 1.38e-08 ***
## day 3.21549 0.60543 1204.86569 5.311 1.30e-07 ***
## event 7.64697 0.95286 1204.86531 8.025 2.39e-15 ***
## ryff 0.20032 0.06896 41.61795 2.905 0.00586 **
## event:ryff -0.09039 0.01341 1204.66271 -6.741 2.43e-11 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff
## instance 0.045
## day -0.068 -0.945
## event -0.394 0.016 -0.017
## ryff -0.973 0.002 -0.003 0.384
## event:ryff 0.385 -0.012 0.014 -0.978 -0.392intmodel1_pos <- lmer(positive_score ~ 1 + instance + day + event +
ryff + ryff:event + (1|id) + (0 + instance|id) + (0 + event|id),
data= mlm)
summary(intmodel1_pos)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: positive_score ~ 1 + instance + day + event + ryff + ryff:event +
## (1 | id) + (0 + instance | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 7788.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.9956 -0.5836 -0.0055 0.5517 4.1257
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 31.87258 5.6456
## id.1 instance 0.06167 0.2483
## id.2 event 6.38992 2.5278
## Residual 26.12131 5.1109
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 4.80321 5.23257 29.44857 0.918 0.36611
## instance -0.66535 0.11307 412.48288 -5.885 8.27e-09 ***
## day 3.32080 0.54965 1136.03257 6.042 2.06e-09 ***
## event 7.07678 2.30569 29.93978 3.069 0.00453 **
## ryff 0.20480 0.07425 28.89020 2.758 0.00998 **
## event:ryff -0.08491 0.03275 29.55394 -2.592 0.01467 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event ryff
## instance 0.027
## day -0.051 -0.870
## event -0.135 0.007 -0.008
## ryff -0.975 0.009 -0.008 0.130
## event:ryff 0.130 -0.005 0.007 -0.978 -0.131anova(ex4model_pos,intmodel1_pos)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## ex4model_pos: positive_score ~ 1 + instance + day + event + ryff + (1 | id) + (0 + instance | id) + (0 + event | id)
## intmodel1_pos: positive_score ~ 1 + instance + day + event + ryff + ryff:event + (1 | id) + (0 + instance | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## ex4model_pos 9 7803.9 7850.0 -3892.9 7785.9
## intmodel1_pos 10 7799.4 7850.6 -3889.7 7779.4 6.5174 1 0.01068 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1crosslevel4_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_sd:instance + (1|id), data = mlm)## Warning: Some predictor variables are on very different scales: consider
## rescaling
## Warning: Some predictor variables are on very different scales: consider
## rescalingsummary(crosslevel4_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_sd:instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6757.9
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9945 -0.5305 -0.1410 0.3768 6.2211
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.065 2.840
## Residual 12.404 3.522
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.337e+01 9.475e-01 2.431e+02 14.109 < 2e-16 ***
## instance -3.336e-01 8.243e-02 1.201e+03 -4.047 5.52e-05 ***
## day 1.886e+00 3.774e-01 1.201e+03 4.998 6.66e-07 ***
## event -8.353e-01 2.476e-01 1.214e+03 -3.374 0.000765 ***
## neg_sd 1.416e-01 4.187e-02 1.211e+03 3.381 0.000744 ***
## neg_ac -8.968e+00 3.516e+00 8.572e+02 -2.551 0.010922 *
## neg_mssd 3.326e-02 1.072e-02 1.231e+03 3.101 0.001971 **
## neg_pac 1.961e+00 2.010e+00 1.231e+03 0.976 0.329428
## instance:neg_sd -6.130e-03 3.526e-03 1.201e+03 -1.738 0.082388 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance -0.055
## day -0.212 -0.816
## event -0.115 -0.006 0.011
## neg_sd -0.227 0.351 -0.011 -0.630
## neg_ac 0.528 -0.011 0.031 0.165 -0.117
## neg_mssd -0.111 0.011 -0.003 0.047 0.020 0.120
## neg_pac -0.342 -0.001 0.001 0.011 -0.090 -0.019 0.035
## instnc:ng_s 0.345 -0.500 -0.007 -0.001 -0.676 -0.022 -0.006 -0.001
## fit warnings:
## Some predictor variables are on very different scales: consider rescalingcrosslevel5_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_sd:day + (1|id), data = mlm)
summary(crosslevel5_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_sd:day + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6752.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0148 -0.5347 -0.1481 0.3765 6.1561
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.067 2.840
## Residual 12.382 3.519
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.94489 0.99041 282.63139 13.070 < 2e-16 ***
## instance -0.40491 0.07134 1201.27954 -5.676 1.73e-08 ***
## day 2.38023 0.43632 1200.77361 5.455 5.93e-08 ***
## event -0.83637 0.24737 1214.39172 -3.381 0.000745 ***
## neg_sd 0.17735 0.04849 1208.58268 3.657 0.000266 ***
## neg_ac -9.01305 3.51253 858.01145 -2.566 0.010458 *
## neg_mssd 0.03327 0.01071 1230.96723 3.105 0.001944 **
## neg_pac 1.95656 2.00858 1230.67756 0.974 0.330198
## day:neg_sd -0.04227 0.01862 1200.63587 -2.270 0.023368 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance 0.128
## day -0.395 -0.816
## event -0.109 -0.007 0.009
## neg_sd -0.335 0.014 0.376 -0.545
## neg_ac 0.507 -0.026 0.032 0.165 -0.105
## neg_mssd -0.106 0.009 0.001 0.047 0.018 0.120
## neg_pac -0.326 -0.002 0.001 0.011 -0.078 -0.019 0.035
## day:neg_sd 0.441 -0.002 -0.503 0.002 -0.772 -0.011 -0.006 0.001crosslevel6_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_ac:instance + (1|id), data = mlm)
summary(crosslevel6_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_ac:instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6748.1
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9610 -0.5424 -0.1498 0.3807 6.2717
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.031 2.834
## Residual 12.405 3.522
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 14.82891 1.02563 306.58330 14.458 < 2e-16 ***
## instance -0.52799 0.10036 1201.74484 -5.261 1.69e-07 ***
## day 1.89774 0.37756 1201.09435 5.026 5.76e-07 ***
## event -0.82884 0.24762 1214.58072 -3.347 0.000841 ***
## neg_sd 0.09246 0.03087 1218.68948 2.995 0.002797 **
## neg_ac -2.88513 5.00744 1111.88974 -0.576 0.564618
## neg_mssd 0.03318 0.01072 1230.97688 3.094 0.002016 **
## neg_pac 1.98685 2.01036 1230.63523 0.988 0.323195
## instance:neg_ac -0.81874 0.47008 1203.51713 -1.742 0.081819 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance -0.262
## day -0.182 -0.689
## event -0.098 -0.016 0.012
## neg_sd 0.008 0.014 -0.022 -0.855
## neg_ac 0.703 -0.513 0.039 0.127 -0.125
## neg_mssd -0.099 0.005 -0.003 0.047 0.022 0.086
## neg_pac -0.311 -0.007 0.002 0.011 -0.122 -0.008 0.035
## instnc:ng_c -0.499 0.703 -0.024 -0.016 -0.001 -0.712 -0.003 -0.008crosslevel7_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_ac:day + (1|id), data = mlm)
summary(crosslevel7_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_ac:day + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6746.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9527 -0.5406 -0.1485 0.3829 6.2646
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.045 2.836
## Residual 12.427 3.525
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 14.66942 1.18909 463.27508 12.337 < 2e-16 ***
## instance -0.40788 0.07153 1201.28882 -5.702 1.49e-08 ***
## day 1.53126 0.53438 1203.11613 2.865 0.00424 **
## event -0.83061 0.24787 1214.66413 -3.351 0.00083 ***
## neg_sd 0.09199 0.03090 1218.84541 2.977 0.00297 **
## neg_ac -4.00973 6.51479 1194.34348 -0.615 0.53835
## neg_mssd 0.03323 0.01073 1230.97349 3.096 0.00200 **
## neg_pac 1.95950 2.01206 1230.63531 0.974 0.33031
## day:neg_ac -2.50378 2.69836 1205.79657 -0.928 0.35365
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance 0.080
## day -0.588 -0.639
## event -0.077 -0.008 -0.007
## neg_sd -0.003 0.021 -0.005 -0.856
## neg_ac 0.789 -0.048 -0.583 0.108 -0.109
## neg_mssd -0.080 0.009 -0.009 0.047 0.022 0.073
## neg_pac -0.272 -0.002 0.000 0.011 -0.122 -0.010 0.035
## day:neg_ac -0.664 0.041 0.707 -0.022 0.015 -0.842 -0.010 -0.001crosslevel8_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_mssd:instance + (1|id), data = mlm)## Warning: Some predictor variables are on very different scales: consider
## rescaling
## Warning: Some predictor variables are on very different scales: consider
## rescalingsummary(crosslevel8_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_mssd:instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6755.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9692 -0.5442 -0.1443 0.3684 6.3200
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.023 2.832
## Residual 12.372 3.517
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.446e+01 9.119e-01 2.074e+02 15.859 < 2e-16 ***
## instance -4.584e-01 7.438e-02 1.201e+03 -6.163 9.71e-10 ***
## day 1.889e+00 3.770e-01 1.201e+03 5.010 6.25e-07 ***
## event -8.341e-01 2.473e-01 1.214e+03 -3.373 0.000766 ***
## neg_sd 9.224e-02 3.083e-02 1.219e+03 2.992 0.002825 **
## neg_ac -8.459e+00 3.519e+00 8.436e+02 -2.404 0.016443 *
## neg_mssd -4.006e-03 1.822e-02 1.224e+03 -0.220 0.826064
## neg_pac 1.840e+00 2.008e+00 1.231e+03 0.916 0.359593
## instance:neg_mssd 4.865e-03 1.931e-03 1.210e+03 2.519 0.011898 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance 0.069
## day -0.216 -0.908
## event -0.118 -0.008 0.011
## neg_sd 0.008 0.020 -0.022 -0.855
## neg_ac 0.571 -0.045 0.032 0.165 -0.179
## neg_mssd -0.251 0.235 -0.007 0.025 0.015 0.012
## neg_pac -0.360 0.005 0.001 0.010 -0.122 -0.021 0.039
## instnc:ng_m 0.228 -0.284 0.007 0.002 -0.002 0.072 -0.809 -0.023
## fit warnings:
## Some predictor variables are on very different scales: consider rescalingcrosslevel9_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_mssd:day + (1|id), data = mlm)
summary(crosslevel9_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_mssd:day + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6755.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9609 -0.5417 -0.1470 0.3692 6.3187
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.03 2.834
## Residual 12.40 3.522
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.439e+01 9.248e-01 2.176e+02 15.560 < 2e-16 ***
## instance -4.059e-01 7.141e-02 1.201e+03 -5.685 1.64e-08 ***
## day 1.691e+00 3.926e-01 1.202e+03 4.306 1.79e-05 ***
## event -8.363e-01 2.476e-01 1.214e+03 -3.378 0.000754 ***
## neg_sd 9.242e-02 3.087e-02 1.219e+03 2.994 0.002809 **
## neg_ac -8.742e+00 3.520e+00 8.472e+02 -2.484 0.013198 *
## neg_mssd -2.457e-03 2.281e-02 1.218e+03 -0.108 0.914249
## neg_pac 1.893e+00 2.011e+00 1.231e+03 0.942 0.346537
## day:neg_mssd 1.830e-02 1.035e-02 1.207e+03 1.768 0.077363 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance 0.137
## day -0.283 -0.907
## event -0.118 -0.007 0.011
## neg_sd 0.009 0.020 -0.021 -0.855
## neg_ac 0.563 -0.026 0.014 0.165 -0.179
## neg_mssd -0.297 0.009 0.242 0.023 0.010 0.006
## neg_pac -0.355 -0.002 0.006 0.011 -0.122 -0.020 0.032
## day:ng_mssd 0.277 -0.006 -0.275 -0.002 0.000 0.057 -0.883 -0.018crosslevel10_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_pac:instance + (1|id), data = mlm)
summary(crosslevel10_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_pac:instance + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6745.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0067 -0.5342 -0.1447 0.3891 6.2623
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.055 2.838
## Residual 12.371 3.517
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.76085 1.00420 299.12864 12.708 < 2e-16 ***
## instance -0.25151 0.09396 1201.28698 -2.677 0.007537 **
## day 1.87211 0.37695 1201.08325 4.966 7.8e-07 ***
## event -0.83862 0.24726 1214.38863 -3.392 0.000717 ***
## neg_sd 0.09288 0.03083 1218.61274 3.013 0.002642 **
## neg_ac -8.94599 3.51112 857.43492 -2.548 0.011011 *
## neg_mssd 0.03339 0.01071 1230.96733 3.117 0.001867 **
## neg_pac 8.97535 3.44047 1217.79357 2.609 0.009198 **
## instance:neg_pac -0.88592 0.35278 1200.75902 -2.511 0.012160 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance -0.207
## day -0.193 -0.724
## event -0.106 -0.009 0.011
## neg_sd 0.005 0.019 -0.022 -0.855
## neg_ac 0.496 -0.008 0.031 0.165 -0.179
## neg_mssd -0.107 0.013 -0.003 0.047 0.022 0.120
## neg_pac -0.566 0.528 -0.008 0.002 -0.066 0.003 0.028
## instnc:ng_p 0.466 -0.651 0.010 0.005 -0.006 -0.018 -0.009 -0.812crosslevel11_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_pac +
neg_pac:day + (1|id), data = mlm)
summary(crosslevel11_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_pac + neg_pac:day + (1 | id)
## Data: mlm
##
## REML criterion at convergence: 6742.3
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -3.0103 -0.5337 -0.1496 0.3860 6.2470
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 8.053 2.838
## Residual 12.374 3.518
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.37688 1.09237 390.58501 11.330 < 2e-16 ***
## instance -0.40156 0.07133 1201.29129 -5.629 2.25e-08 ***
## day 2.65656 0.49167 1200.81896 5.403 7.89e-08 ***
## event -0.83809 0.24728 1214.39514 -3.389 0.000724 ***
## neg_sd 0.09282 0.03083 1218.62041 3.011 0.002661 **
## neg_ac -9.00561 3.51113 857.39198 -2.565 0.010490 *
## neg_mssd 0.03336 0.01071 1230.96830 3.115 0.001884 **
## neg_pac 11.20977 4.27107 1212.26296 2.625 0.008785 **
## day:neg_pac -4.64179 1.89150 1200.67696 -2.454 0.014268 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss neg_pc
## instance 0.105
## day -0.513 -0.711
## event -0.097 -0.007 0.006
## neg_sd 0.004 0.020 -0.013 -0.855
## neg_ac 0.458 -0.025 0.031 0.165 -0.179
## neg_mssd -0.099 0.009 0.003 0.047 0.022 0.120
## neg_pac -0.653 0.017 0.567 0.001 -0.053 0.001 0.024
## day:neg_pac 0.582 -0.021 -0.642 0.004 -0.006 -0.011 -0.008 -0.883#add significant interactions
intmodel1_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_sd:day + neg_mssd:day +
+ neg_pac:instance + neg_pac:day + (1|id) +
(0 + instance|id) + (0 + day|id) + (0 + event|id),
data= mlm)
summary(intmodel1_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_sd:day + neg_mssd:day + +neg_pac:instance +
## neg_pac:day + (1 | id) + (0 + instance | id) + (0 + day |
## id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6665.4
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.8702 -0.5089 -0.1307 0.3504 5.6821
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 7.045189 2.65428
## id.1 instance 0.002467 0.04967
## id.2 day 0.847794 0.92076
## id.3 event 1.850278 1.36025
## Residual 10.632598 3.26077
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.290e+01 1.007e+00 1.766e+02 12.812 < 2e-16 ***
## instance -1.021e-01 1.727e-01 5.514e+02 -0.591 0.554700
## day 6.660e-01 8.392e-01 9.606e+02 0.794 0.427648
## event -1.108e+00 3.396e-01 6.425e+01 -3.263 0.001769 **
## neg_sd 1.834e-01 4.594e-02 1.175e+03 3.992 6.96e-05 ***
## neg_ac -1.587e+01 4.312e+00 4.472e+02 -3.680 0.000262 ***
## neg_mssd 6.991e-04 2.365e-02 8.476e+02 0.030 0.976421
## day:neg_sd -3.982e-02 1.772e-02 1.152e+03 -2.247 0.024805 *
## day:neg_mssd 1.699e-02 1.098e-02 7.610e+02 1.548 0.122002
## instance:neg_pac -1.829e+00 9.375e-01 6.550e+02 -1.951 0.051531 .
## day:neg_pac 9.274e+00 4.110e+00 8.649e+02 2.257 0.024282 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss dy:ng_s dy:ng_m
## instance 0.079
## day -0.251 -0.902
## event -0.013 -0.012 0.015
## neg_sd -0.395 -0.064 0.254 -0.387
## neg_ac 0.614 0.005 0.003 0.180 -0.152
## neg_mssd -0.308 0.034 0.113 0.015 0.072 0.003
## day:neg_sd 0.440 0.074 -0.306 0.005 -0.768 0.011 -0.082
## day:ng_mssd 0.306 -0.026 -0.142 0.000 -0.066 0.068 -0.878 0.094
## instnc:ng_p -0.036 -0.922 0.816 0.013 0.075 -0.009 -0.032 -0.080 0.023
## day:neg_pac 0.027 0.890 -0.836 -0.018 -0.071 -0.004 0.034 0.053 -0.023
## inst:_
## instance
## day
## event
## neg_sd
## neg_ac
## neg_mssd
## day:neg_sd
## day:ng_mssd
## instnc:ng_p
## day:neg_pac -0.967anova(ex2model_neg, intmodel1_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## ex2model_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_pac + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## intmodel1_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_sd:day + neg_mssd:day + +neg_pac:instance + neg_pac:day + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## ex2model_neg 13 6674.9 6741.5 -3324.5 6648.9
## intmodel1_neg 16 6674.8 6756.7 -3321.4 6642.8 6.1576 3 0.1042#only significant
intmodel2_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd +
neg_sd:day + neg_pac:instance + neg_pac:day +
(1|id) + (0 + instance|id) + (0 + day|id) + (0 + event|id),
data= mlm)
summary(intmodel2_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_sd:day + neg_pac:instance + neg_pac:day +
## (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6660.6
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.9078 -0.5212 -0.1372 0.3475 5.6861
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 7.247271 2.69208
## id.1 instance 0.004655 0.06823
## id.2 day 0.829043 0.91052
## id.3 event 1.873975 1.36893
## Residual 10.625327 3.25965
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 12.43608 0.96325 160.03732 12.910 < 2e-16 ***
## instance -0.09378 0.17349 569.11782 -0.541 0.58904
## day 0.84287 0.83238 1010.54874 1.013 0.31150
## event -1.10895 0.34067 63.43081 -3.255 0.00182 **
## neg_sd 0.18778 0.04584 1168.36049 4.096 4.49e-05 ***
## neg_ac -16.28637 4.31213 456.85357 -3.777 0.00018 ***
## neg_mssd 0.03287 0.01133 931.37766 2.902 0.00379 **
## day:neg_sd -0.04224 0.01764 1141.98377 -2.395 0.01677 *
## instance:neg_pac -1.86726 0.94112 683.31014 -1.984 0.04765 *
## day:neg_pac 9.44004 4.12154 891.60207 2.290 0.02223 *
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss dy:ng_s inst:_
## instance 0.091
## day -0.219 -0.915
## event -0.013 -0.012 0.016
## neg_sd -0.393 -0.066 0.248 -0.387
## neg_ac 0.623 0.007 0.012 0.180 -0.149
## neg_mssd -0.086 0.025 -0.024 0.032 0.031 0.131
## day:neg_sd 0.432 0.077 -0.297 0.006 -0.767 0.005 0.001
## instnc:ng_p -0.045 -0.921 0.829 0.013 0.077 -0.011 -0.024 -0.083
## day:neg_pac 0.036 0.890 -0.849 -0.018 -0.073 -0.002 0.028 0.056 -0.967anova(ex2model_neg, intmodel2_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## ex2model_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_pac + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## intmodel2_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_sd:day + neg_pac:instance + neg_pac:day + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## ex2model_neg 13 6674.9 6741.5 -3324.5 6648.9
## intmodel2_neg 15 6675.2 6752.0 -3322.6 6645.2 3.7046 2 0.1569#only mssd
intmodel3_neg <- lmer(negative_score ~ 1 + instance + day + event +
neg_sd + neg_ac + neg_mssd + neg_mssd:day +
(1|id) + (0 + instance|id) + (0 + day|id) + (0 + event|id),
data= mlm)
summary(intmodel3_neg)## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac +
## neg_mssd + neg_mssd:day + (1 | id) + (0 + instance | id) +
## (0 + day | id) + (0 + event | id)
## Data: mlm
##
## REML criterion at convergence: 6673.8
##
## Scaled residuals:
## Min 1Q Median 3Q Max
## -2.6829 -0.5233 -0.1181 0.3557 5.7594
##
## Random effects:
## Groups Name Variance Std.Dev.
## id (Intercept) 7.291017 2.70019
## id.1 instance 0.002552 0.05051
## id.2 day 0.825339 0.90848
## id.3 event 1.661083 1.28883
## Residual 10.728271 3.27540
## Number of obs: 1240, groups: id, 32
##
## Fixed effects:
## Estimate Std. Error df t value Pr(>|t|)
## (Intercept) 1.393e+01 9.081e-01 1.216e+02 15.339 < 2e-16 ***
## instance -4.112e-01 6.728e-02 2.972e+02 -6.111 3.10e-09 ***
## day 1.727e+00 4.048e-01 3.722e+02 4.265 2.54e-05 ***
## event -1.073e+00 3.311e-01 7.229e+01 -3.242 0.001797 **
## neg_sd 1.087e-01 2.931e-02 1.182e+03 3.708 0.000219 ***
## neg_ac -1.517e+01 4.289e+00 4.390e+02 -3.536 0.000449 ***
## neg_mssd -5.485e-03 2.364e-02 8.550e+02 -0.232 0.816553
## day:neg_mssd 1.992e-02 1.096e-02 7.662e+02 1.818 0.069417 .
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Correlation of Fixed Effects:
## (Intr) instnc day event neg_sd neg_ac ng_mss
## instance 0.132
## day -0.260 -0.813
## event -0.020 -0.001 0.005
## neg_sd -0.095 0.016 -0.021 -0.622
## neg_ac 0.675 -0.011 0.006 0.182 -0.230
## neg_mssd -0.304 0.013 0.249 0.016 0.017 0.003
## day:ng_mssd 0.295 -0.011 -0.284 -0.001 0.010 0.068 -0.878anova(ex2model_neg, intmodel3_neg)## refitting model(s) with ML (instead of REML)## Data: mlm
## Models:
## ex2model_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_pac + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## intmodel3_neg: negative_score ~ 1 + instance + day + event + neg_sd + neg_ac + neg_mssd + neg_mssd:day + (1 | id) + (0 + instance | id) + (0 + day | id) + (0 + event | id)
## npar AIC BIC logLik deviance Chisq Df Pr(>Chisq)
## ex2model_neg 13 6674.9 6741.5 -3324.5 6648.9
## intmodel3_neg 13 6679.7 6746.3 -3326.8 6653.7 0 0